#### Power supply adjustments

Joe Subich, W4TV

On 2019-11-07 4:53 PM, Rob Sherwood wrote:
Ok, it isn't a straight line, but what I was trying to imply was it consistently improved with higher voltage.
OK. I just wanted to make it clear that running below 12V caused a
massive increase in IMD while at 14.4-14.5V one was entering a
region of diminishing returns.

Since the transistors (at least in the early PA) are rated for
16V breakdown (collector to emitter), I would be leery of pushing
the voltage any higher than 15 V in receive.

You also measure "best case" IMD by using a short/heavy cable.
A 5' long #12 cable has about 5 times the resistance/voltage drop
of your test configuration. The 5' cable is more than 3+ times
the length of your 16" test cable and resistance of #12 is 50%
greater than #10 for the same length ( 3 x 1.5 = 4.5).

73,

... Joe, W4TV

On 2019-11-07 4:53 PM, Rob Sherwood wrote:
Ok, it isn't a straight line, but what I was trying to imply was it consistently improved with higher voltage.
It didn't plateau and then get worse.
Approximately a 10 dB difference from 10.5 to 14.8 volts, and approximately 3 d difference between 13.8 and 14.8 volts.
Rob, NC0B
-----Original Message-----
From: Elecraft-K3@groups.io [mailto:Elecraft-K3@groups.io] On Behalf Of Joe Subich, W4TV
Sent: Thursday, November 07, 2019 1:50 PM
To: Elecraft-K3@groups.io
Subject: Re: [Elecraft-K3] Power supply adjustments
On 2019-11-07 2:50 PM, Rob Sherwood wrote:

The pattern I measured was linear, the higher the voltage, the lower
> the third-order IMD.
Given the rate of change, I would not call the slope linear.
There is a terrible 2.8 dB worsening of IMD when the voltage drops from
12.0 to 11.5 volts. In addition, you change from 0.5V steps below 13.0
V to 0.4V steps above which hides an improvement in dB/V between 13.0
and 14.6V (1.0, 1.1, 1.3, 1.3 dB/0.4V). The improvement "flattens out"
again above 14.6V (only 0.3 dB/.02V).
It does make one wonder if the K3/K3S could achieve -36 dB/PEP at
50-60 W with a 14.5V supply <G> (particularly if the LPA were set up
for better IMD at that level).
73,
... Joe, W4TV
On 2019-11-07 2:50 PM, Rob Sherwood wrote:
This morning I decided to measure a K3 in respect to third-order IMD vs. power supply voltage as read on the K3 metering. The pattern I measured was linear, the higher the voltage, the lower the third-order IMD. At really low voltages the much higher odd-order products degrade significantly. Test was on 20 meters, at a nominal 100 watts as per the K3 power setting. Two equal tones were generated within the K3. Note: at 14.8 volts on TX, the voltage on RX was 15.1 volts. At 13.8 volts on TX, the voltage on RX was 14.0. The power supply was an Astron VS-35M, and the length of PowerPole cable was 16 inches.

Voltage IMD in dBc Note: Add 6 dB for PEP method
11.5 -14.1 dBc
12.0 -16.9 dBc
12.5 -17.9 dBc
13.0 -19.0 dBc
13.4 -20.0 dBc
13.8 -21.1 dBc
14.2 -22.4 dBc
14.6 -23.7 dBc
14.8 -24.0 dBc

Rob, NC0B

Rob Sherwood

Hi Joe,

Absolutely correct about the cable issue. The first cable I started using is 6 feet long, and was purchased at HRO as a pre-built cable. The maximum voltage I could get at the K3, measured on its LCD, was 14.6 volts. Going to the 16 inch cable pushed the reading to 14.8 volts.

With the 16 inch cable, the Astron measures 13.995 volts on an HP 3456 DVM. The K3 displays 13.8 volts on the LCD.

I do consider a 3 dB improvement from 13.8 to 14.8 significant, if not earth shattering.

Rob, NC0B

-----Original Message-----
From: Elecraft-K3@groups.io [mailto:Elecraft-K3@groups.io] On Behalf Of Joe Subich, W4TV
Sent: Thursday, November 07, 2019 3:48 PM
To: Elecraft-K3@groups.io
Subject: Re: [Elecraft-K3] Power supply adjustments

On 2019-11-07 4:53 PM, Rob Sherwood wrote:
Ok, it isn't a straight line, but what I was trying to imply was it
> consistently improved with higher voltage.

OK. I just wanted to make it clear that running below 12V caused a
massive increase in IMD while at 14.4-14.5V one was entering a
region of diminishing returns.

Since the transistors (at least in the early PA) are rated for
16V breakdown (collector to emitter), I would be leery of pushing
the voltage any higher than 15 V in receive.

You also measure "best case" IMD by using a short/heavy cable.
A 5' long #12 cable has about 5 times the resistance/voltage drop
of your test configuration. The 5' cable is more than 3+ times
the length of your 16" test cable and resistance of #12 is 50%
greater than #10 for the same length ( 3 x 1.5 = 4.5).

73,

... Joe, W4TV

On 2019-11-07 4:53 PM, Rob Sherwood wrote:
Ok, it isn't a straight line, but what I was trying to imply was it consistently improved with higher voltage.
It didn't plateau and then get worse.
Approximately a 10 dB difference from 10.5 to 14.8 volts, and approximately 3 d difference between 13.8 and 14.8 volts.
Rob, NC0B

-----Original Message-----
From: Elecraft-K3@groups.io [mailto:Elecraft-K3@groups.io] On Behalf Of Joe Subich, W4TV
Sent: Thursday, November 07, 2019 1:50 PM
To: Elecraft-K3@groups.io
Subject: Re: [Elecraft-K3] Power supply adjustments

On 2019-11-07 2:50 PM, Rob Sherwood wrote:

The pattern I measured was linear, the higher the voltage, the lower
> the third-order IMD.

Given the rate of change, I would not call the slope linear.

There is a terrible 2.8 dB worsening of IMD when the voltage drops from
12.0 to 11.5 volts. In addition, you change from 0.5V steps below 13.0
V to 0.4V steps above which hides an improvement in dB/V between 13.0
and 14.6V (1.0, 1.1, 1.3, 1.3 dB/0.4V). The improvement "flattens out"
again above 14.6V (only 0.3 dB/.02V).

It does make one wonder if the K3/K3S could achieve -36 dB/PEP at
50-60 W with a 14.5V supply <G> (particularly if the LPA were set up
for better IMD at that level).

73,

... Joe, W4TV

On 2019-11-07 2:50 PM, Rob Sherwood wrote:
This morning I decided to measure a K3 in respect to third-order IMD vs. power supply voltage as read on the K3 metering. The pattern I measured was linear, the higher the voltage, the lower the third-order IMD. At really low voltages the much higher odd-order products degrade significantly. Test was on 20 meters, at a nominal 100 watts as per the K3 power setting. Two equal tones were generated within the K3. Note: at 14.8 volts on TX, the voltage on RX was 15.1 volts. At 13.8 volts on TX, the voltage on RX was 14.0. The power supply was an Astron VS-35M, and the length of PowerPole cable was 16 inches.

Voltage IMD in dBc Note: Add 6 dB for PEP method
11.5 -14.1 dBc
12.0 -16.9 dBc
12.5 -17.9 dBc
13.0 -19.0 dBc
13.4 -20.0 dBc
13.8 -21.1 dBc
14.2 -22.4 dBc
14.6 -23.7 dBc
14.8 -24.0 dBc

Rob, NC0B

dalej

Thanks Rob, looks like 35 watts is the sweet spot and where I drive my KPA500 plus or minus a little.

Dale j.

On 07, Nov 2019, at 16:46, Rob Sherwood <rob@...> wrote:

In general with factory supplied power cables, particularly very long Icom cables, the voltage drop in the cable is 0.5 volts, if not significantly more. I cut my Icom cables in half, which also removes the fuses and fuse holders and their associated poor connection issues.

My Astron power supplies are over voltage and over current protected. An Icom PS-125 or PS-126 doesn't have any fuses in the short power lead, so I feel comfortable removing the fuses. My TS-890S runs on an Astron RS-35M with a shortened cable with no fuses.

In any case, as asked for, here is some more data all taken at 13.8 volts in transmit mode as read by the K3 LCD.
Also the PA in this K3 is Rev C.

Wattage 3rd order IMD
100 watts -21.1 dBc
75 watts -27.8 dBc
50 watts -33.7 dBc
35 watts -36.7 dBc
12 watts -23.4 dBc
10 watts -29.2 dBc

Note at 12 watts and below the driver is providing the RF output. The PA is out of the circuit.

Rob, NC0B

-----Original Message-----
From: Elecraft-K3@groups.io [mailto:Elecraft-K3@groups.io] On Behalf Of Bill K9YEQ
Sent: Thursday, November 07, 2019 2:56 PM
To: Elecraft-K3@groups.io
Subject: Re: [Elecraft-K3] Power supply adjustments

That is significant. I checked my voltage at the K3S and need to make some changes real soon. I generally only run < 30 watts into my KPA1500, but need to improve voltage at the radio for those few times I run barefoot at 60+ watts.

73,
Bill
K9YEQ

https://wrj-tech.com/

-----Original Message-----
From: Elecraft-K3@groups.io <Elecraft-K3@groups.io> On Behalf Of Rob Sherwood
Sent: Thursday, November 7, 2019 3:54 PM
To: 'Elecraft-K3@groups.io' <Elecraft-K3@groups.io>
Subject: Re: [Elecraft-K3] Power supply adjustments

Ok, it isn't a straight line, but what I was trying to imply was it consistently improved with higher voltage.
It didn't plateau and then get worse.
Approximately a 10 dB difference from 10.5 to 14.8 volts, and approximately 3 d difference between 13.8 and 14.8 volts.
Rob, NC0B

-----Original Message-----
From: Elecraft-K3@groups.io [mailto:Elecraft-K3@groups.io] On Behalf Of Joe Subich, W4TV
Sent: Thursday, November 07, 2019 1:50 PM
To: Elecraft-K3@groups.io
Subject: Re: [Elecraft-K3] Power supply adjustments

On 2019-11-07 2:50 PM, Rob Sherwood wrote:

The pattern I measured was linear, the higher the voltage, the lower
the third-order IMD.
Given the rate of change, I would not call the slope linear.

There is a terrible 2.8 dB worsening of IMD when the voltage drops from
12.0 to 11.5 volts. In addition, you change from 0.5V steps below 13.0
V to 0.4V steps above which hides an improvement in dB/V between 13.0
and 14.6V (1.0, 1.1, 1.3, 1.3 dB/0.4V). The improvement "flattens out"
again above 14.6V (only 0.3 dB/.02V).

It does make one wonder if the K3/K3S could achieve -36 dB/PEP at
50-60 W with a 14.5V supply <G> (particularly if the LPA were set up
for better IMD at that level).

73,

... Joe, W4TV

On 2019-11-07 2:50 PM, Rob Sherwood wrote:
This morning I decided to measure a K3 in respect to third-order IMD vs. power supply voltage as read on the K3 metering. The pattern I measured was linear, the higher the voltage, the lower the third-order IMD. At really low voltages the much higher odd-order products degrade significantly. Test was on 20 meters, at a nominal 100 watts as per the K3 power setting. Two equal tones were generated within the K3. Note: at 14.8 volts on TX, the voltage on RX was 15.1 volts. At 13.8 volts on TX, the voltage on RX was 14.0. The power supply was an Astron VS-35M, and the length of PowerPole cable was 16 inches.

Voltage IMD in dBc Note: Add 6 dB for PEP method
11.5 -14.1 dBc
12.0 -16.9 dBc
12.5 -17.9 dBc
13.0 -19.0 dBc
13.4 -20.0 dBc
13.8 -21.1 dBc
14.2 -22.4 dBc
14.6 -23.7 dBc
14.8 -24.0 dBc

Rob, NC0B

Richard Smith

This has been a very helpful set of emails.   I adjusted my RS-35M to 14.5V.  The K3S reading was 14.1V.  When I transmitted the K3S dropped to 11.9V. Dang. Got rid of Anderson power strip, shorted and connected K3S power direct to power supply.  K3S now reads 14.3V and drops to 13.9V at transmit.  Thanks for all the info.

KU4NZ

On Nov 7, 2019, at 4:55 PM, Rob Sherwood <rob@...> wrote:

Hi Joe,

Absolutely correct about the cable issue.  The first cable I started using is 6 feet long, and was purchased at HRO as a pre-built cable.  The maximum voltage I could get at the K3, measured on its LCD, was 14.6 volts.  Going to the 16 inch cable pushed the reading to 14.8 volts.

With the 16 inch cable, the Astron measures 13.995 volts on an HP 3456 DVM.  The K3 displays 13.8 volts on the LCD.

I do consider a 3 dB improvement from 13.8 to 14.8 significant, if not earth shattering.

Rob, NC0B

-----Original Message-----
From: Elecraft-K3@groups.io [mailto:Elecraft-K3@groups.io] On Behalf Of Joe Subich, W4TV
Sent: Thursday, November 07, 2019 3:48 PM
To: Elecraft-K3@groups.io
Subject: Re: [Elecraft-K3] Power supply adjustments

On 2019-11-07 4:53 PM, Rob Sherwood wrote:
Ok, it isn't a straight line, but what I was trying to imply was it
consistently improved with higher voltage.

OK. I just wanted to make it clear that running below 12V caused a
massive increase in IMD while at 14.4-14.5V one was entering a
region of diminishing returns.

Since the transistors (at least in the early PA) are rated for
16V breakdown (collector to emitter), I would be leery of pushing
the voltage any higher than 15 V in receive.

You also measure "best case" IMD by using a short/heavy cable.
A 5' long #12 cable has about 5 times the resistance/voltage drop
of your test configuration.  The 5' cable is more than 3+ times
the length of your 16" test cable and resistance of #12 is 50%
greater than #10 for the same length ( 3 x 1.5 = 4.5).

73,

... Joe, W4TV

On 2019-11-07 4:53 PM, Rob Sherwood wrote:
Ok, it isn't a straight line, but what I was trying to imply was it consistently improved with higher voltage.
It didn't plateau and then get worse.
Approximately a 10 dB difference from 10.5 to 14.8 volts, and approximately 3 d difference between 13.8 and 14.8 volts.
Rob, NC0B

-----Original Message-----
From: Elecraft-K3@groups.io [mailto:Elecraft-K3@groups.io] On Behalf Of Joe Subich, W4TV
Sent: Thursday, November 07, 2019 1:50 PM
To: Elecraft-K3@groups.io
Subject: Re: [Elecraft-K3] Power supply adjustments

On 2019-11-07 2:50 PM, Rob Sherwood wrote:

The pattern I measured was linear, the higher the voltage, the lower
the third-order IMD.

Given the rate of change, I would not call the slope linear.

There is a terrible 2.8 dB worsening of IMD when the voltage drops from
12.0 to 11.5 volts.  In addition, you change from 0.5V steps below 13.0
V to 0.4V steps above which hides an improvement in dB/V between 13.0
and 14.6V (1.0, 1.1, 1.3, 1.3 dB/0.4V).  The improvement "flattens out"
again above 14.6V (only 0.3 dB/.02V).

It does make one wonder if the K3/K3S could achieve -36 dB/PEP at
50-60 W with a 14.5V supply <G> (particularly if the LPA were set up
for better IMD at that level).

73,

... Joe, W4TV

On 2019-11-07 2:50 PM, Rob Sherwood wrote:
This morning I decided to measure a K3 in respect to third-order IMD vs. power supply voltage as read on the K3 metering.  The pattern I measured was linear, the higher the voltage, the lower the third-order IMD.  At really low voltages the much higher odd-order products degrade significantly.  Test was on 20 meters, at a nominal 100 watts as per the K3 power setting.  Two equal tones were generated within the K3.  Note: at 14.8 volts on TX, the voltage on RX was 15.1 volts.  At 13.8 volts on TX, the voltage on RX was 14.0. The power supply was an Astron VS-35M, and the length of PowerPole cable was 16 inches.

Voltage                IMD in dBc   Note: Add 6 dB for PEP method
11.5                        -14.1 dBc
12.0                        -16.9 dBc
12.5                        -17.9 dBc
13.0                        -19.0 dBc
13.4                        -20.0 dBc
13.8                        -21.1 dBc
14.2                        -22.4 dBc
14.6                        -23.7 dBc
14.8                        -24.0 dBc

Rob, NC0B

Joe Subich, W4TV

On 2019-11-07 5:55 PM, Rob Sherwood wrote:
Absolutely correct about the cable issue. The first cable I started using is 6 feet long, and was purchased at HRO as a pre-built cable. The maximum voltage I could get at the K3, measured on its LCD, was 14.6 volts. Going to the 16 inch cable pushed the reading to 14.8 volts.
Yes, it's a trade off between the ability to access the transceiver
easily (pull it forward for service) and voltage drop. Going to #10
instead of the usual #12 helps a good bit and keeping the cable less
than 30" (2 1/2 feet) is probably a good compromise.

I do consider a 3 dB improvement from 13.8 to 14.8 significant, if
not earth shattering.
Yes, it's significant but I'd be just as happy with 2.5 dB and a bit
more breakdown margin for the transistors - particularly with the
original PA. If the higher voltage is coupled with a shorter, heavier
power cable the total benefit is probably closer to 5 dB.

73,

... Joe, W4TV

On 2019-11-07 5:55 PM, Rob Sherwood wrote:
Hi Joe,
Absolutely correct about the cable issue. The first cable I started using is 6 feet long, and was purchased at HRO as a pre-built cable. The maximum voltage I could get at the K3, measured on its LCD, was 14.6 volts. Going to the 16 inch cable pushed the reading to 14.8 volts.
With the 16 inch cable, the Astron measures 13.995 volts on an HP 3456 DVM. The K3 displays 13.8 volts on the LCD.
I do consider a 3 dB improvement from 13.8 to 14.8 significant, if not earth shattering.
Rob, NC0B
-----Original Message-----
From: Elecraft-K3@groups.io [mailto:Elecraft-K3@groups.io] On Behalf Of Joe Subich, W4TV
Sent: Thursday, November 07, 2019 3:48 PM
To: Elecraft-K3@groups.io
Subject: Re: [Elecraft-K3] Power supply adjustments
On 2019-11-07 4:53 PM, Rob Sherwood wrote:
Ok, it isn't a straight line, but what I was trying to imply was it
> consistently improved with higher voltage.
OK. I just wanted to make it clear that running below 12V caused a
massive increase in IMD while at 14.4-14.5V one was entering a
region of diminishing returns.
Since the transistors (at least in the early PA) are rated for
16V breakdown (collector to emitter), I would be leery of pushing
the voltage any higher than 15 V in receive.
You also measure "best case" IMD by using a short/heavy cable.
A 5' long #12 cable has about 5 times the resistance/voltage drop
of your test configuration. The 5' cable is more than 3+ times
the length of your 16" test cable and resistance of #12 is 50%
greater than #10 for the same length ( 3 x 1.5 = 4.5).
73,
... Joe, W4TV
On 2019-11-07 4:53 PM, Rob Sherwood wrote:
Ok, it isn't a straight line, but what I was trying to imply was it consistently improved with higher voltage.
It didn't plateau and then get worse.
Approximately a 10 dB difference from 10.5 to 14.8 volts, and approximately 3 d difference between 13.8 and 14.8 volts.
Rob, NC0B

-----Original Message-----
From: Elecraft-K3@groups.io [mailto:Elecraft-K3@groups.io] On Behalf Of Joe Subich, W4TV
Sent: Thursday, November 07, 2019 1:50 PM
To: Elecraft-K3@groups.io
Subject: Re: [Elecraft-K3] Power supply adjustments

On 2019-11-07 2:50 PM, Rob Sherwood wrote:

The pattern I measured was linear, the higher the voltage, the lower
> the third-order IMD.

Given the rate of change, I would not call the slope linear.

There is a terrible 2.8 dB worsening of IMD when the voltage drops from
12.0 to 11.5 volts. In addition, you change from 0.5V steps below 13.0
V to 0.4V steps above which hides an improvement in dB/V between 13.0
and 14.6V (1.0, 1.1, 1.3, 1.3 dB/0.4V). The improvement "flattens out"
again above 14.6V (only 0.3 dB/.02V).

It does make one wonder if the K3/K3S could achieve -36 dB/PEP at
50-60 W with a 14.5V supply <G> (particularly if the LPA were set up
for better IMD at that level).

73,

... Joe, W4TV

On 2019-11-07 2:50 PM, Rob Sherwood wrote:
This morning I decided to measure a K3 in respect to third-order IMD vs. power supply voltage as read on the K3 metering. The pattern I measured was linear, the higher the voltage, the lower the third-order IMD. At really low voltages the much higher odd-order products degrade significantly. Test was on 20 meters, at a nominal 100 watts as per the K3 power setting. Two equal tones were generated within the K3. Note: at 14.8 volts on TX, the voltage on RX was 15.1 volts. At 13.8 volts on TX, the voltage on RX was 14.0. The power supply was an Astron VS-35M, and the length of PowerPole cable was 16 inches.

Voltage IMD in dBc Note: Add 6 dB for PEP method
11.5 -14.1 dBc
12.0 -16.9 dBc
12.5 -17.9 dBc
13.0 -19.0 dBc
13.4 -20.0 dBc
13.8 -21.1 dBc
14.2 -22.4 dBc
14.6 -23.7 dBc
14.8 -24.0 dBc

Rob, NC0B

Rob Sherwood

Wayne and I had an off-reflector chat about my numbers which were meant to show how important the actual voltage at the back of the radio is.  I did clearly say that my numbers were dBc, not PEP, and to add 6 dB if you wanted the PEP values.  This distinction may well not be obvious to most readers.

In the “good old days” all transmit intermodulation data was published correctly in dBc.  Then at some point “number inflation” came along and the OEMs and the ARRL decided to reference intermodulation to PEP (peak envelope power) by adding 6 dB to all the numbers.  Of course this “slight of hand” didn’t affect the distortion on the air, it simply made the published numbers larger.

Older folks like myself (72 years old) remember the WATT wars in the Hi Fi world of the 1970s for power amplifiers.  All sorts of imaginary numbers were invented instead of using true RMS power into an 8 ohm or 4 ohm load.  I digress.

The ITU spec, which isn’t required for amateur service is -25 dBc, or 31 dB PEP.  That is the MINIMUM considered acceptable in the commercial world. If you wan to compare my numbers to what appears in QST, add 6 dB.  Again, the absolute number wasn’t the point of my data, which was answering a question as to whether the voltage actually reaching the back of the radio was significant.  Obviously Richard, KU4NZ, (below) found the discussion useful.

Just a 6 foot PowerPole cable purchased from HRO has a 0.5 volt drop in my test setup at a 100 watt 2-tone test.  This is rather typical of a relatively good cable with a practical length.  Some OEM cables are as long as 10 feet, and some at least in the past had very small wire (Ten-Tec).

So there we have it.  Hopefully we all now know that dBc used correctly isn’t the same as PEP, and to consider all the sources of voltage drop in your DC wiring.

73, Rob, NC0B

From: Elecraft-K3@groups.io [mailto:Elecraft-K3@groups.io] On Behalf Of Richard Smith
Sent: Thursday, November 07, 2019 5:31 PM
To: Elecraft-K3@groups.io
Subject: Re: [Elecraft-K3] Power supply adjustments

This has been a very helpful set of emails.   I adjusted my RS-35M to 14.5V.  The K3S reading was 14.1V.  When I transmitted the K3S dropped to 11.9V. Dang. Got rid of Anderson power strip, shorted and connected K3S power direct to power supply.  K3S now reads 14.3V and drops to 13.9V at transmit.  Thanks for all the info.

KU4NZ

On Nov 7, 2019, at 4:55 PM, Rob Sherwood <rob@...> wrote:

Hi Joe,

Absolutely correct about the cable issue.  The first cable I started using is 6 feet long, and was purchased at HRO as a pre-built cable.  The maximum voltage I could get at the K3, measured on its LCD, was 14.6 volts.  Going to the 16 inch cable pushed the reading to 14.8 volts.

With the 16 inch cable, the Astron measures 13.995 volts on an HP 3456 DVM.  The K3 displays 13.8 volts on the LCD.

I do consider a 3 dB improvement from 13.8 to 14.8 significant, if not earth shattering.

Rob, NC0B

-----Original Message-----
From:
Elecraft-K3@groups.io [mailto:Elecraft-K3@groups.io] On Behalf Of Joe Subich, W4TV
Sent: Thursday, November 07, 2019 3:48 PM
To:
Elecraft-K3@groups.io
Subject: Re: [Elecraft-K3] Power supply adjustments

On 2019-11-07 4:53 PM, Rob Sherwood wrote:

Ok, it isn't a straight line, but what I was trying to imply was it
consistently improved with higher voltage.

OK. I just wanted to make it clear that running below 12V caused a
massive increase in IMD while at 14.4-14.5V one was entering a
region of diminishing returns.

Since the transistors (at least in the early PA) are rated for
16V breakdown (collector to emitter), I would be leery of pushing
the voltage any higher than 15 V in receive.

You also measure "best case" IMD by using a short/heavy cable.
A 5' long #12 cable has about 5 times the resistance/voltage drop
of your test configuration.  The 5' cable is more than 3+ times
the length of your 16" test cable and resistance of #12 is 50%
greater than #10 for the same length ( 3 x 1.5 = 4.5).

73,

... Joe, W4TV

On 2019-11-07 4:53 PM, Rob Sherwood wrote:

Ok, it isn't a straight line, but what I was trying to imply was it consistently improved with higher voltage.
It didn't plateau and then get worse.
Approximately a 10 dB difference from 10.5 to 14.8 volts, and approximately 3 d difference between 13.8 and 14.8 volts.
Rob, NC0B

-----Original Message-----
From: Elecraft-K3@groups.io [mailto:Elecraft-K3@groups.io] On Behalf Of Joe Subich, W4TV
Sent: Thursday, November 07, 2019 1:50 PM
To: Elecraft-K3@groups.io
Subject: Re: [Elecraft-K3] Power supply adjustments

On 2019-11-07 2:50 PM, Rob Sherwood wrote:

The pattern I measured was linear, the higher the voltage, the lower
the third-order IMD.

Given the rate of change, I would not call the slope linear.

There is a terrible 2.8 dB worsening of IMD when the voltage drops from
12.0 to 11.5 volts.  In addition, you change from 0.5V steps below 13.0
V to 0.4V steps above which hides an improvement in dB/V between 13.0
and 14.6V (1.0, 1.1, 1.3, 1.3 dB/0.4V).  The improvement "flattens out"
again above 14.6V (only 0.3 dB/.02V).

It does make one wonder if the K3/K3S could achieve -36 dB/PEP at
50-60 W with a 14.5V supply <G> (particularly if the LPA were set up
for better IMD at that level).

73,

... Joe, W4TV

On 2019-11-07 2:50 PM, Rob Sherwood wrote:

This morning I decided to measure a K3 in respect to third-order IMD vs. power supply voltage as read on the K3 metering.  The pattern I measured was linear, the higher the voltage, the lower the third-order IMD.  At really low voltages the much higher odd-order products degrade significantly.  Test was on 20 meters, at a nominal 100 watts as per the K3 power setting.  Two equal tones were generated within the K3.  Note: at 14.8 volts on TX, the voltage on RX was 15.1 volts.  At 13.8 volts on TX, the voltage on RX was 14.0. The power supply was an Astron VS-35M, and the length of PowerPole cable was 16 inches.

Voltage                IMD in dBc   Note: Add 6 dB for PEP method
11.5                        -14.1 dBc
12.0                        -16.9 dBc
12.5                        -17.9 dBc
13.0                        -19.0 dBc
13.4                        -20.0 dBc
13.8                        -21.1 dBc
14.2                        -22.4 dBc
14.6                        -23.7 dBc
14.8                        -24.0 dBc

Rob, NC0B

Larry

If 35 Watts is the cleanest, does that mean that running it at 35W in
the digital modes is preferred

or acceptable?

Larry

W0OGH

On 11/7/2019 2:50 PM, Rob Sherwood wrote:
There is no question the K3 is better at 35 watts. I can make some measurements at different power levels. Rob, NC0B

-----Original Message-----
From: Elecraft-K3@groups.io [mailto:Elecraft-K3@groups.io] On Behalf Of N6KR
Sent: Thursday, November 07, 2019 1:42 PM
To: Elecraft-K3@groups.io
Subject: Re: [Elecraft-K3] Power supply adjustments

On average, IMD will be much better at 35 W than at 100 W. And as with all class AB amps using MOSFETs, the IMD will vary per-band.

Wayne
N6KR

On Nov 7, 2019, at 12:37 PM, dalej via Groups.Io <dalej2=mac.com@groups.io> wrote:

wondering what the IMD is with power set to 35 watts and the PS set to 13.8

Dale k9vuj

On 07, Nov 2019, at 13:50, Rob Sherwood <rob@...> wrote:

This morning I decided to measure a K3 in respect to third-order IMD vs. power supply voltage as read on the K3 metering. The pattern I measured was linear, the higher the voltage, the lower the third-order IMD. At really low voltages the much higher odd-order products degrade significantly. Test was on 20 meters, at a nominal 100 watts as per the K3 power setting. Two equal tones were generated within the K3. Note: at 14.8 volts on TX, the voltage on RX was 15.1 volts. At 13.8 volts on TX, the voltage on RX was 14.0. The power supply was an Astron VS-35M, and the length of PowerPole cable was 16 inches.

Voltage IMD in dBc Note: Add 6 dB for PEP method
11.5 -14.1 dBc
12.0 -16.9 dBc
12.5 -17.9 dBc
13.0 -19.0 dBc
13.4 -20.0 dBc
13.8 -21.1 dBc
14.2 -22.4 dBc
14.6 -23.7 dBc
14.8 -24.0 dBc

Rob, NC0B

From: Elecraft-K3@groups.io [mailto:Elecraft-K3@groups.io] On Behalf Of Dennis Ashworth
Sent: Thursday, November 07, 2019 11:28 AM
To: Elecraft-K3@groups.io
Subject: Re: [Elecraft-K3] Power supply adjustments

Are the supply voltage differences powering the K3 really material to IMD numbers? I’ve not seen data to answer the question either way ... just curious.
Dennis, K7FL

On Nov 7, 2019, at 10:22 AM, Michael Mickelson <@KD8DZ> wrote:

﻿
Bob, thanks for that information. I have two of the 30Amp Astron supplies and will make an adjustment to a higher setting. The discussion on the effects of supply voltage on the K3/K3S tranSmitter performance has been very inStructive.

Thanks to all who contributed. Mike KD8DZ

On Wed, Nov 6, 2019 at 5:18 PM Bob McGraw - K4TAX <rmcgraw@...> wrote:
Why not get Alexa to adjust the voltage for you? "Alexa, turn the
supply on and set the voltage for 14.5."

The supplies are usually in the category of the "set and forget" type.
Thus the caution of "no user serviceable parts inside".

I adjusted one of mine this afternoon. I put it on the desk, removed
the 4 screws holding then cover, plugged in the cord and turned the
supply on. I picked up my light in one hand and my plastic adjustment
tool in the other {you do have and use a plastic adjustment tool don't
you?} and made the adjustment. Nowhere at any time did my hands or
fingers come close to any voltage. I did put my Fluke DMV on the
output terminals first to monitor the voltage. At 14.49 indicated, I
stopped. Turned the supply off, put the cover on and secured it with
the 4 screws.

73

Bob, K4TAX

On 11/6/2019 3:54 PM, Rick Bates, NK7I wrote:
And for some stupid reason, when they could have put that pot on the
underside of the circuit board (which ends up on the top) for simpler
access, they didn't. So you're reaching in around open hot high
voltage OR turning it on/off a lot of times to get it set right. Plan
B could have been to put the pot on the edge or board flush with a
hole for a plastic tool to access...

I maintain that if the design engineers (of ANY field) EVER had to
work on, fix, adjust their product, the design would be far different
and techs would bless them instead of using 'adult language'
curses... ;-)

I hold that Elecraft 'gets' this, their designs are modular and except
cable routing and some panel connecting, is boringly simple.

Rick NK7I

On 11/6/2019 10:32 AM, Rob Sherwood wrote:
The Astron power supplies I own have an internal pot to set the
voltage. I would adjust it with nothing connected. The can likely go
over 16 volts. I have 3 RS-35Ms, but have never used a 50. For some
reason removing the top covers are a pain. Rob, NC0B

--
Michael Mickelson
J. Reid Anderson Professor of Physics and Astronomy, emeritus
Denison University
740-587-0687 H
740-644-1238 M

Rob Sherwood

Hi Larry,

WSJT X does not have multiple tones transmitted at the same time. You can run a class C amplifier with JT9, for example, a mode I use on 630 meters. There may well be other modes that use multiple simultaneous tones, in which case linearity would matter. Hopefully others will chime in who have more knowledge of all things digital.

Rob, NC0B

-----Original Message-----
From: Elecraft-K3@groups.io [mailto:Elecraft-K3@groups.io] On Behalf Of Larry
Sent: Thursday, November 07, 2019 7:08 PM
To: Elecraft-K3@groups.io
Subject: Re: [Elecraft-K3] Power supply adjustments

If 35 Watts is the cleanest, does that mean that running it at 35W in
the digital modes is preferred

or acceptable?

Larry

W0OGH

On 11/7/2019 2:50 PM, Rob Sherwood wrote:
There is no question the K3 is better at 35 watts. I can make some measurements at different power levels. Rob, NC0B

-----Original Message-----
From: Elecraft-K3@groups.io [mailto:Elecraft-K3@groups.io] On Behalf Of N6KR
Sent: Thursday, November 07, 2019 1:42 PM
To: Elecraft-K3@groups.io
Subject: Re: [Elecraft-K3] Power supply adjustments

On average, IMD will be much better at 35 W than at 100 W. And as with all class AB amps using MOSFETs, the IMD will vary per-band.

Wayne
N6KR

On Nov 7, 2019, at 12:37 PM, dalej via Groups.Io <dalej2=mac.com@groups.io> wrote:

wondering what the IMD is with power set to 35 watts and the PS set to 13.8

Dale k9vuj

On 07, Nov 2019, at 13:50, Rob Sherwood <rob@...> wrote:

This morning I decided to measure a K3 in respect to third-order IMD vs. power supply voltage as read on the K3 metering. The pattern I measured was linear, the higher the voltage, the lower the third-order IMD. At really low voltages the much higher odd-order products degrade significantly. Test was on 20 meters, at a nominal 100 watts as per the K3 power setting. Two equal tones were generated within the K3. Note: at 14.8 volts on TX, the voltage on RX was 15.1 volts. At 13.8 volts on TX, the voltage on RX was 14.0. The power supply was an Astron VS-35M, and the length of PowerPole cable was 16 inches.

Voltage IMD in dBc Note: Add 6 dB for PEP method
11.5 -14.1 dBc
12.0 -16.9 dBc
12.5 -17.9 dBc
13.0 -19.0 dBc
13.4 -20.0 dBc
13.8 -21.1 dBc
14.2 -22.4 dBc
14.6 -23.7 dBc
14.8 -24.0 dBc

Rob, NC0B

From: Elecraft-K3@groups.io [mailto:Elecraft-K3@groups.io] On Behalf Of Dennis Ashworth
Sent: Thursday, November 07, 2019 11:28 AM
To: Elecraft-K3@groups.io
Subject: Re: [Elecraft-K3] Power supply adjustments

Are the supply voltage differences powering the K3 really material to IMD numbers? I’ve not seen data to answer the question either way ... just curious.
Dennis, K7FL

On Nov 7, 2019, at 10:22 AM, Michael Mickelson <@KD8DZ> wrote:

﻿
Bob, thanks for that information. I have two of the 30Amp Astron supplies and will make an adjustment to a higher setting. The discussion on the effects of supply voltage on the K3/K3S tranSmitter performance has been very inStructive.

Thanks to all who contributed. Mike KD8DZ

On Wed, Nov 6, 2019 at 5:18 PM Bob McGraw - K4TAX <rmcgraw@...> wrote:
Why not get Alexa to adjust the voltage for you? "Alexa, turn the
supply on and set the voltage for 14.5."

The supplies are usually in the category of the "set and forget" type.
Thus the caution of "no user serviceable parts inside".

I adjusted one of mine this afternoon. I put it on the desk, removed
the 4 screws holding then cover, plugged in the cord and turned the
supply on. I picked up my light in one hand and my plastic adjustment
tool in the other {you do have and use a plastic adjustment tool don't
you?} and made the adjustment. Nowhere at any time did my hands or
fingers come close to any voltage. I did put my Fluke DMV on the
output terminals first to monitor the voltage. At 14.49 indicated, I
stopped. Turned the supply off, put the cover on and secured it with
the 4 screws.

73

Bob, K4TAX

On 11/6/2019 3:54 PM, Rick Bates, NK7I wrote:
And for some stupid reason, when they could have put that pot on the
underside of the circuit board (which ends up on the top) for simpler
access, they didn't. So you're reaching in around open hot high
voltage OR turning it on/off a lot of times to get it set right. Plan
B could have been to put the pot on the edge or board flush with a
hole for a plastic tool to access...

I maintain that if the design engineers (of ANY field) EVER had to
work on, fix, adjust their product, the design would be far different
and techs would bless them instead of using 'adult language'
curses... ;-)

I hold that Elecraft 'gets' this, their designs are modular and except
cable routing and some panel connecting, is boringly simple.

Rick NK7I

On 11/6/2019 10:32 AM, Rob Sherwood wrote:
The Astron power supplies I own have an internal pot to set the
voltage. I would adjust it with nothing connected. The can likely go
over 16 volts. I have 3 RS-35Ms, but have never used a 50. For some
reason removing the top covers are a pain. Rob, NC0B

--
Michael Mickelson
J. Reid Anderson Professor of Physics and Astronomy, emeritus
Denison University
740-587-0687 H
740-644-1238 M

Joe Subich, W4TV

On 2019-11-07 9:45 PM, Rob Sherwood wrote:

WSJT X does not have multiple tones transmitted at the same time. You
can run a class C amplifier with JT9
Except Joe Taylor, et al are adapting the GFSK modulation (wave shaping
for minimum occupied bandwidth) to the older modes. The gaussian filter
adds some AM component which will require more linear amplification in
order to maintain the minimum transmitted bandwidth.

I don't know that running any of the WSJT-X modes (perhaps with the
exception of MSK144 which is OQPSK) through a class C amplifier will
make the GFSK signals massively wider but they will be somewhat more
broad.

73,

... Joe, W4TV

On 2019-11-07 9:45 PM, Rob Sherwood wrote:
Hi Larry,
WSJT X does not have multiple tones transmitted at the same time. You can run a class C amplifier with JT9, for example, a mode I use on 630 meters. There may well be other modes that use multiple simultaneous tones, in which case linearity would matter. Hopefully others will chime in who have more knowledge of all things digital.
Rob, NC0B
-----Original Message-----
From: Elecraft-K3@groups.io [mailto:Elecraft-K3@groups.io] On Behalf Of Larry
Sent: Thursday, November 07, 2019 7:08 PM
To: Elecraft-K3@groups.io
Subject: Re: [Elecraft-K3] Power supply adjustments
If 35 Watts is the cleanest, does that mean that running it at 35W in
the digital modes is preferred
or acceptable?
Larry
W0OGH
On 11/7/2019 2:50 PM, Rob Sherwood wrote:
There is no question the K3 is better at 35 watts. I can make some measurements at different power levels. Rob, NC0B

-----Original Message-----
From: Elecraft-K3@groups.io [mailto:Elecraft-K3@groups.io] On Behalf Of N6KR
Sent: Thursday, November 07, 2019 1:42 PM
To: Elecraft-K3@groups.io
Subject: Re: [Elecraft-K3] Power supply adjustments

On average, IMD will be much better at 35 W than at 100 W. And as with all class AB amps using MOSFETs, the IMD will vary per-band.

Wayne
N6KR

On Nov 7, 2019, at 12:37 PM, dalej via Groups.Io <dalej2=mac.com@groups.io> wrote:

wondering what the IMD is with power set to 35 watts and the PS set to 13.8

Dale k9vuj

On 07, Nov 2019, at 13:50, Rob Sherwood <rob@...> wrote:

This morning I decided to measure a K3 in respect to third-order IMD vs. power supply voltage as read on the K3 metering. The pattern I measured was linear, the higher the voltage, the lower the third-order IMD. At really low voltages the much higher odd-order products degrade significantly. Test was on 20 meters, at a nominal 100 watts as per the K3 power setting. Two equal tones were generated within the K3. Note: at 14.8 volts on TX, the voltage on RX was 15.1 volts. At 13.8 volts on TX, the voltage on RX was 14.0. The power supply was an Astron VS-35M, and the length of PowerPole cable was 16 inches.

Voltage IMD in dBc Note: Add 6 dB for PEP method
11.5 -14.1 dBc
12.0 -16.9 dBc
12.5 -17.9 dBc
13.0 -19.0 dBc
13.4 -20.0 dBc
13.8 -21.1 dBc
14.2 -22.4 dBc
14.6 -23.7 dBc
14.8 -24.0 dBc

Rob, NC0B

From: Elecraft-K3@groups.io [mailto:Elecraft-K3@groups.io] On Behalf Of Dennis Ashworth
Sent: Thursday, November 07, 2019 11:28 AM
To: Elecraft-K3@groups.io
Subject: Re: [Elecraft-K3] Power supply adjustments

Are the supply voltage differences powering the K3 really material to IMD numbers? I’ve not seen data to answer the question either way ... just curious.
Dennis, K7FL

On Nov 7, 2019, at 10:22 AM, Michael Mickelson <@KD8DZ> wrote:

﻿
Bob, thanks for that information. I have two of the 30Amp Astron supplies and will make an adjustment to a higher setting. The discussion on the effects of supply voltage on the K3/K3S tranSmitter performance has been very inStructive.

Thanks to all who contributed. Mike KD8DZ

On Wed, Nov 6, 2019 at 5:18 PM Bob McGraw - K4TAX <rmcgraw@...> wrote:
Why not get Alexa to adjust the voltage for you? "Alexa, turn the
supply on and set the voltage for 14.5."

The supplies are usually in the category of the "set and forget" type.
Thus the caution of "no user serviceable parts inside".

I adjusted one of mine this afternoon. I put it on the desk, removed
the 4 screws holding then cover, plugged in the cord and turned the
supply on. I picked up my light in one hand and my plastic adjustment
tool in the other {you do have and use a plastic adjustment tool don't
you?} and made the adjustment. Nowhere at any time did my hands or
fingers come close to any voltage. I did put my Fluke DMV on the
output terminals first to monitor the voltage. At 14.49 indicated, I
stopped. Turned the supply off, put the cover on and secured it with
the 4 screws.

73

Bob, K4TAX

On 11/6/2019 3:54 PM, Rick Bates, NK7I wrote:
And for some stupid reason, when they could have put that pot on the
underside of the circuit board (which ends up on the top) for simpler
access, they didn't. So you're reaching in around open hot high
voltage OR turning it on/off a lot of times to get it set right. Plan
B could have been to put the pot on the edge or board flush with a
hole for a plastic tool to access...

I maintain that if the design engineers (of ANY field) EVER had to
work on, fix, adjust their product, the design would be far different
and techs would bless them instead of using 'adult language'
curses... ;-)

I hold that Elecraft 'gets' this, their designs are modular and except
cable routing and some panel connecting, is boringly simple.

Rick NK7I

On 11/6/2019 10:32 AM, Rob Sherwood wrote:
The Astron power supplies I own have an internal pot to set the
voltage. I would adjust it with nothing connected. The can likely go
over 16 volts. I have 3 RS-35Ms, but have never used a 50. For some
reason removing the top covers are a pain. Rob, NC0B

--
Michael Mickelson
J. Reid Anderson Professor of Physics and Astronomy, emeritus
Denison University
740-587-0687 H
740-644-1238 M

Rob Sherwood

Hi Joe,

I thought Joe Taylor explained that the tone that was jumping in frequency does so less abruptly in the latest release than before. Why would a smoother frequency change be degraded by a non-linear amplifier?

Rob, NC0B

On Nov 7, 2019, at 8:13 PM, Joe Subich, W4TV <lists@...> wrote:

﻿On 2019-11-07 9:45 PM, Rob Sherwood wrote:

WSJT X does not have multiple tones transmitted at the same time. You
can run a class C amplifier with JT9
Except Joe Taylor, et al are adapting the GFSK modulation (wave shaping
for minimum occupied bandwidth) to the older modes. The gaussian filter
adds some AM component which will require more linear amplification in
order to maintain the minimum transmitted bandwidth.

I don't know that running any of the WSJT-X modes (perhaps with the
exception of MSK144 which is OQPSK) through a class C amplifier will
make the GFSK signals massively wider but they will be somewhat more
broad.

73,

... Joe, W4TV

On 2019-11-07 9:45 PM, Rob Sherwood wrote:
Hi Larry,
WSJT X does not have multiple tones transmitted at the same time. You can run a class C amplifier with JT9, for example, a mode I use on 630 meters. There may well be other modes that use multiple simultaneous tones, in which case linearity would matter. Hopefully others will chime in who have more knowledge of all things digital.
Rob, NC0B
-----Original Message-----
From: Elecraft-K3@groups.io [mailto:Elecraft-K3@groups.io] On Behalf Of Larry
Sent: Thursday, November 07, 2019 7:08 PM
To: Elecraft-K3@groups.io
Subject: Re: [Elecraft-K3] Power supply adjustments
If 35 Watts is the cleanest, does that mean that running it at 35W in
the digital modes is preferred
or acceptable?
Larry
W0OGH
On 11/7/2019 2:50 PM, Rob Sherwood wrote:
There is no question the K3 is better at 35 watts. I can make some measurements at different power levels. Rob, NC0B

-----Original Message-----
From: Elecraft-K3@groups.io [mailto:Elecraft-K3@groups.io] On Behalf Of N6KR
Sent: Thursday, November 07, 2019 1:42 PM
To: Elecraft-K3@groups.io
Subject: Re: [Elecraft-K3] Power supply adjustments

On average, IMD will be much better at 35 W than at 100 W. And as with all class AB amps using MOSFETs, the IMD will vary per-band.

Wayne
N6KR

On Nov 7, 2019, at 12:37 PM, dalej via Groups.Io <dalej2=mac.com@groups.io> wrote:

wondering what the IMD is with power set to 35 watts and the PS set to 13.8

Dale k9vuj

On 07, Nov 2019, at 13:50, Rob Sherwood <rob@...> wrote:

This morning I decided to measure a K3 in respect to third-order IMD vs. power supply voltage as read on the K3 metering. The pattern I measured was linear, the higher the voltage, the lower the third-order IMD. At really low voltages the much higher odd-order products degrade significantly. Test was on 20 meters, at a nominal 100 watts as per the K3 power setting. Two equal tones were generated within the K3. Note: at 14.8 volts on TX, the voltage on RX was 15.1 volts. At 13.8 volts on TX, the voltage on RX was 14.0. The power supply was an Astron VS-35M, and the length of PowerPole cable was 16 inches.

Voltage IMD in dBc Note: Add 6 dB for PEP method
11.5 -14.1 dBc
12.0 -16.9 dBc
12.5 -17.9 dBc
13.0 -19.0 dBc
13.4 -20.0 dBc
13.8 -21.1 dBc
14.2 -22.4 dBc
14.6 -23.7 dBc
14.8 -24.0 dBc

Rob, NC0B

From: Elecraft-K3@groups.io [mailto:Elecraft-K3@groups.io] On Behalf Of Dennis Ashworth
Sent: Thursday, November 07, 2019 11:28 AM
To: Elecraft-K3@groups.io
Subject: Re: [Elecraft-K3] Power supply adjustments

Are the supply voltage differences powering the K3 really material to IMD numbers? I’ve not seen data to answer the question either way ... just curious.
Dennis, K7FL

On Nov 7, 2019, at 10:22 AM, Michael Mickelson <@KD8DZ> wrote:

﻿
Bob, thanks for that information. I have two of the 30Amp Astron supplies and will make an adjustment to a higher setting. The discussion on the effects of supply voltage on the K3/K3S tranSmitter performance has been very inStructive.

Thanks to all who contributed. Mike KD8DZ

On Wed, Nov 6, 2019 at 5:18 PM Bob McGraw - K4TAX <rmcgraw@...> wrote:
Why not get Alexa to adjust the voltage for you? "Alexa, turn the
supply on and set the voltage for 14.5."

The supplies are usually in the category of the "set and forget" type.
Thus the caution of "no user serviceable parts inside".

I adjusted one of mine this afternoon. I put it on the desk, removed
the 4 screws holding then cover, plugged in the cord and turned the
supply on. I picked up my light in one hand and my plastic adjustment
tool in the other {you do have and use a plastic adjustment tool don't
you?} and made the adjustment. Nowhere at any time did my hands or
fingers come close to any voltage. I did put my Fluke DMV on the
output terminals first to monitor the voltage. At 14.49 indicated, I
stopped. Turned the supply off, put the cover on and secured it with
the 4 screws.

73

Bob, K4TAX

On 11/6/2019 3:54 PM, Rick Bates, NK7I wrote:
And for some stupid reason, when they could have put that pot on the
underside of the circuit board (which ends up on the top) for simpler
access, they didn't. So you're reaching in around open hot high
voltage OR turning it on/off a lot of times to get it set right. Plan
B could have been to put the pot on the edge or board flush with a
hole for a plastic tool to access...

I maintain that if the design engineers (of ANY field) EVER had to
work on, fix, adjust their product, the design would be far different
and techs would bless them instead of using 'adult language'
curses... ;-)

I hold that Elecraft 'gets' this, their designs are modular and except
cable routing and some panel connecting, is boringly simple.

Rick NK7I

On 11/6/2019 10:32 AM, Rob Sherwood wrote:
The Astron power supplies I own have an internal pot to set the
voltage. I would adjust it with nothing connected. The can likely go
over 16 volts. I have 3 RS-35Ms, but have never used a 50. For some
reason removing the top covers are a pain. Rob, NC0B

--
Michael Mickelson
J. Reid Anderson Professor of Physics and Astronomy, emeritus
Denison University
740-587-0687 H
740-644-1238 M

W1MD CFL.rr.com

Thank you Rob. Most informative discussion regarding DC loss and impact on ‘on air’ performance for ‘any’ radio.

40 years a ham (sadly I’m still one of the young ones at 55) and still things to learn every day.

Marty
W1MD

On Nov 7, 2019, at 8:26 PM, Rob Sherwood <rob@...> wrote:

﻿

Wayne and I had an off-reflector chat about my numbers which were meant to show how important the actual voltage at the back of the radio is.  I did clearly say that my numbers were dBc, not PEP, and to add 6 dB if you wanted the PEP values.  This distinction may well not be obvious to most readers.

In the “good old days” all transmit intermodulation data was published correctly in dBc.  Then at some point “number inflation” came along and the OEMs and the ARRL decided to reference intermodulation to PEP (peak envelope power) by adding 6 dB to all the numbers.  Of course this “slight of hand” didn’t affect the distortion on the air, it simply made the published numbers larger.

Older folks like myself (72 years old) remember the WATT wars in the Hi Fi world of the 1970s for power amplifiers.  All sorts of imaginary numbers were invented instead of using true RMS power into an 8 ohm or 4 ohm load.  I digress.

The ITU spec, which isn’t required for amateur service is -25 dBc, or 31 dB PEP.  That is the MINIMUM considered acceptable in the commercial world. If you wan to compare my numbers to what appears in QST, add 6 dB.  Again, the absolute number wasn’t the point of my data, which was answering a question as to whether the voltage actually reaching the back of the radio was significant.  Obviously Richard, KU4NZ, (below) found the discussion useful.

Just a 6 foot PowerPole cable purchased from HRO has a 0.5 volt drop in my test setup at a 100 watt 2-tone test.  This is rather typical of a relatively good cable with a practical length.  Some OEM cables are as long as 10 feet, and some at least in the past had very small wire (Ten-Tec).

So there we have it.  Hopefully we all now know that dBc used correctly isn’t the same as PEP, and to consider all the sources of voltage drop in your DC wiring.

73, Rob, NC0B

From: Elecraft-K3@groups.io [mailto:Elecraft-K3@groups.io] On Behalf Of Richard Smith
Sent: Thursday, November 07, 2019 5:31 PM
To: Elecraft-K3@groups.io
Subject: Re: [Elecraft-K3] Power supply adjustments

This has been a very helpful set of emails.   I adjusted my RS-35M to 14.5V.  The K3S reading was 14.1V.  When I transmitted the K3S dropped to 11.9V. Dang. Got rid of Anderson power strip, shorted and connected K3S power direct to power supply.  K3S now reads 14.3V and drops to 13.9V at transmit.  Thanks for all the info.

KU4NZ

On Nov 7, 2019, at 4:55 PM, Rob Sherwood <rob@...> wrote:

Hi Joe,

Absolutely correct about the cable issue.  The first cable I started using is 6 feet long, and was purchased at HRO as a pre-built cable.  The maximum voltage I could get at the K3, measured on its LCD, was 14.6 volts.  Going to the 16 inch cable pushed the reading to 14.8 volts.

With the 16 inch cable, the Astron measures 13.995 volts on an HP 3456 DVM.  The K3 displays 13.8 volts on the LCD.

I do consider a 3 dB improvement from 13.8 to 14.8 significant, if not earth shattering.

Rob, NC0B

-----Original Message-----
From:
Elecraft-K3@groups.io [mailto:Elecraft-K3@groups.io] On Behalf Of Joe Subich, W4TV
Sent: Thursday, November 07, 2019 3:48 PM
To:
Elecraft-K3@groups.io
Subject: Re: [Elecraft-K3] Power supply adjustments

On 2019-11-07 4:53 PM, Rob Sherwood wrote:

Ok, it isn't a straight line, but what I was trying to imply was it
consistently improved with higher voltage.

OK. I just wanted to make it clear that running below 12V caused a
massive increase in IMD while at 14.4-14.5V one was entering a
region of diminishing returns.

Since the transistors (at least in the early PA) are rated for
16V breakdown (collector to emitter), I would be leery of pushing
the voltage any higher than 15 V in receive.

You also measure "best case" IMD by using a short/heavy cable.
A 5' long #12 cable has about 5 times the resistance/voltage drop
of your test configuration.  The 5' cable is more than 3+ times
the length of your 16" test cable and resistance of #12 is 50%
greater than #10 for the same length ( 3 x 1.5 = 4.5).

73,

... Joe, W4TV

On 2019-11-07 4:53 PM, Rob Sherwood wrote:

Ok, it isn't a straight line, but what I was trying to imply was it consistently improved with higher voltage.
It didn't plateau and then get worse.
Approximately a 10 dB difference from 10.5 to 14.8 volts, and approximately 3 d difference between 13.8 and 14.8 volts.
Rob, NC0B

-----Original Message-----
From: Elecraft-K3@groups.io [mailto:Elecraft-K3@groups.io] On Behalf Of Joe Subich, W4TV
Sent: Thursday, November 07, 2019 1:50 PM
To: Elecraft-K3@groups.io
Subject: Re: [Elecraft-K3] Power supply adjustments

On 2019-11-07 2:50 PM, Rob Sherwood wrote:

The pattern I measured was linear, the higher the voltage, the lower
the third-order IMD.

Given the rate of change, I would not call the slope linear.

There is a terrible 2.8 dB worsening of IMD when the voltage drops from
12.0 to 11.5 volts.  In addition, you change from 0.5V steps below 13.0
V to 0.4V steps above which hides an improvement in dB/V between 13.0
and 14.6V (1.0, 1.1, 1.3, 1.3 dB/0.4V).  The improvement "flattens out"
again above 14.6V (only 0.3 dB/.02V).

It does make one wonder if the K3/K3S could achieve -36 dB/PEP at
50-60 W with a 14.5V supply <G> (particularly if the LPA were set up
for better IMD at that level).

73,

... Joe, W4TV

On 2019-11-07 2:50 PM, Rob Sherwood wrote:

This morning I decided to measure a K3 in respect to third-order IMD vs. power supply voltage as read on the K3 metering.  The pattern I measured was linear, the higher the voltage, the lower the third-order IMD.  At really low voltages the much higher odd-order products degrade significantly.  Test was on 20 meters, at a nominal 100 watts as per the K3 power setting.  Two equal tones were generated within the K3.  Note: at 14.8 volts on TX, the voltage on RX was 15.1 volts.  At 13.8 volts on TX, the voltage on RX was 14.0. The power supply was an Astron VS-35M, and the length of PowerPole cable was 16 inches.

Voltage                IMD in dBc   Note: Add 6 dB for PEP method
11.5                        -14.1 dBc
12.0                        -16.9 dBc
12.5                        -17.9 dBc
13.0                        -19.0 dBc
13.4                        -20.0 dBc
13.8                        -21.1 dBc
14.2                        -22.4 dBc
14.6                        -23.7 dBc
14.8                        -24.0 dBc

Rob, NC0B

Dave Cole

Hi Rob,

Does not WSJT-X transmit multiple tones in Fox/Hound mode?

73, and thanks,
Dave (NK7Z)
https://www.nk7z.net
ARRL Volunteer Examiner
ARRL Technical Specialist
ARRL Asst. Director, NW Division, Technical Resources

On 11/7/19 6:45 PM, Rob Sherwood wrote:
Hi Larry,
WSJT X does not have multiple tones transmitted at the same time. You can run a class C amplifier with JT9, for example, a mode I use on 630 meters. There may well be other modes that use multiple simultaneous tones, in which case linearity would matter. Hopefully others will chime in who have more knowledge of all things digital.
Rob, NC0B

Wes

Except when it isn't.  And in my experience, bipolar amps vary with frequency just like FETs.

Wayne points out something that is totally ignored in these IMD discussions, but something I've mentioned several times before in other discussions: IMD varies with frequency, as well as power output, voltage input, load impedance and the phase of the moon; well many not the latter.  So when someone, says, "The IMD at 13.8 volts is -33 dB", they have provided far too little information.

With the proliferation of cheap SDR receivers, TX IMD measurements, while tedious, are within the grasp of many, especially with Elecraft radios that have built-in two-tone generators.  If anyone is interested I can describe the method I use with an RF Space SDR-IQ.

Wes  N7WS

On 11/7/2019 1:41 PM, N6KR wrote:
On average, IMD will be much better at 35 W than at 100 W. And as with all class AB amps using MOSFETs, the IMD will vary per-band.

Wayne
N6KR

Joe Subich, W4TV

On 2019-11-07 10:58 PM, Rob Sherwood wrote:
Why would a smoother frequency change be degraded by a non-linear amplifier?
Why does a class C amplifier generate key clicks on a CW signal with
an unshaped (fast rise/fall) CW signal?

In essence both CW like Elecraft's raised cosine shaping (which is
also used on their FSK, IIRC) and Joe Taylor's GFSK are filtered
to remove the high frequency components of the keying waveform
turning the rectangular wave (for each tone in n-FSK) into a
"rounded" pulse. It is this "rounding" that Joe Taylor calls a
less abrupt jump.

The class C amplifier - which normally operates well into saturation -
"cuts off" the top portion of that shaping making a sharp transition
from the "ramp up" to the steady on (saturated) state. This is also
why certain Yaesu transceivers with excess IF gain and aggressive ALC
are so "clicky" on CW ... the CW wave shaping is lost in the saturated
(IF in that case) amplifiers.

73,

... Joe, W4TV

On 2019-11-07 10:58 PM, Rob Sherwood wrote:
Hi Joe,
I thought Joe Taylor explained that the tone that was jumping in frequency does so less abruptly in the latest release than before. Why would a smoother frequency change be degraded by a non-linear amplifier?
Rob, NC0B
On Nov 7, 2019, at 8:13 PM, Joe Subich, W4TV <lists@...> wrote:

﻿On 2019-11-07 9:45 PM, Rob Sherwood wrote:

WSJT X does not have multiple tones transmitted at the same time. You
can run a class C amplifier with JT9
Except Joe Taylor, et al are adapting the GFSK modulation (wave shaping
for minimum occupied bandwidth) to the older modes. The gaussian filter
adds some AM component which will require more linear amplification in
order to maintain the minimum transmitted bandwidth.

I don't know that running any of the WSJT-X modes (perhaps with the
exception of MSK144 which is OQPSK) through a class C amplifier will
make the GFSK signals massively wider but they will be somewhat more
broad.

73,

... Joe, W4TV

On 2019-11-07 9:45 PM, Rob Sherwood wrote:
Hi Larry,
WSJT X does not have multiple tones transmitted at the same time. You can run a class C amplifier with JT9, for example, a mode I use on 630 meters. There may well be other modes that use multiple simultaneous tones, in which case linearity would matter. Hopefully others will chime in who have more knowledge of all things digital.
Rob, NC0B

Rob Sherwood

Hi Dave,

That is a good point, if there are multiple simultaneous FT8 and FT4 (?) QSOs going on, then yes there would be intermodulation products. Considering the IMD products are only down 20 some dB, that would produce a lot of phantom signals that could decode on the other end. I definitely decode line-related phantom signals on 630m JT9 on strong signals. I have decoded as many as 2 phantom signals on each side of the mail signal.

Rob, NC0B

-----Original Message-----
From: Elecraft-K3@groups.io [mailto:Elecraft-K3@groups.io] On Behalf Of Dave Cole
Sent: Friday, November 08, 2019 4:44 AM
To: Elecraft-K3@groups.io
Subject: Re: [Elecraft-K3] Power supply adjustments

Hi Rob,

Does not WSJT-X transmit multiple tones in Fox/Hound mode?

73, and thanks,
Dave (NK7Z)
https://www.nk7z.net
ARRL Volunteer Examiner
ARRL Technical Specialist
ARRL Asst. Director, NW Division, Technical Resources

On 11/7/19 6:45 PM, Rob Sherwood wrote:
Hi Larry,

WSJT X does not have multiple tones transmitted at the same time. You can run a class C amplifier with JT9, for example, a mode I use on 630 meters. There may well be other modes that use multiple simultaneous tones, in which case linearity would matter. Hopefully others will chime in who have more knowledge of all things digital.

Rob, NC0B

Rob Sherwood

Wes is 100% right. The magnetics (transformers) of a solid-state amp are having to cope with 160 - 10 or 160 - 6 meters, which is a real challenge. I have measured IMD on a real antenna tuned with a Johnson Matchbox, and varied the SWR up to around 2:1. Certainly the IMD changes with the impedance. A 15 dB variation on third-order distortion isn't unheard of from band to band. The third-order product in general is the most variable, compared to 5th order, and it isn't unusual to see 3rd order depressed below the 5th order product on some bands, and particularly at rated output.

The three TS-990S transceivers I measured from 160 - 10 meters, and from 50 to 180 watts, were more consistent than typical. That said, the first sample was defective on 10 meters with terrible IMD. I never found out what was wrong with it when it was returned to Kenwood.

Rob, NC0B

-----Original Message-----
From: Elecraft-K3@groups.io [mailto:Elecraft-K3@groups.io] On Behalf Of Wes
Sent: Friday, November 08, 2019 6:08 AM
To: Elecraft-K3@groups.io
Subject: Re: [Elecraft-K3] Power supply adjustments

Except when it isn't.  And in my experience, bipolar amps vary with frequency
just like FETs.

Wayne points out something that is totally ignored in these IMD discussions, but
something I've mentioned several times before in other discussions: IMD varies
with frequency, as well as power output, voltage input, load impedance and the
phase of the moon; well many not the latter.  So when someone, says, "The IMD at
13.8 volts is -33 dB", they have provided far too little information.

With the proliferation of cheap SDR receivers, TX IMD measurements, while
tedious, are within the grasp of many, especially with Elecraft radios that have
built-in two-tone generators.  If anyone is interested I can describe the method
I use with an RF Space SDR-IQ.

Wes  N7WS

On 11/7/2019 1:41 PM, N6KR wrote:
On average, IMD will be much better at 35 W than at 100 W. And as with all class AB amps using MOSFETs, the IMD will vary per-band.

Wayne
N6KR

Rob Sherwood

Typo fixed. Main not "Mail". Duh. Rob, NC0B

-----Original Message-----
From: Elecraft-K3@groups.io [mailto:Elecraft-K3@groups.io] On Behalf Of Rob Sherwood
Sent: Friday, November 08, 2019 8:25 AM
To: Elecraft-K3@groups.io
Subject: Re: [Elecraft-K3] Power supply adjustments

Hi Dave,

That is a good point, if there are multiple simultaneous FT8 and FT4 (?) QSOs going on, then yes there would be intermodulation products. Considering the IMD products are only down 20 some dB, that would produce a lot of phantom signals that could decode on the other end. I definitely decode line-related phantom signals on 630m JT9 on strong signals. I have decoded as many as 2 phantom signals on each side of the main signal.

Rob, NC0B

-----Original Message-----
From: Elecraft-K3@groups.io [mailto:Elecraft-K3@groups.io] On Behalf Of Dave Cole
Sent: Friday, November 08, 2019 4:44 AM
To: Elecraft-K3@groups.io
Subject: Re: [Elecraft-K3] Power supply adjustments

Hi Rob,

Does not WSJT-X transmit multiple tones in Fox/Hound mode?

73, and thanks,
Dave (NK7Z)
https://www.nk7z.net
ARRL Volunteer Examiner
ARRL Technical Specialist
ARRL Asst. Director, NW Division, Technical Resources

On 11/7/19 6:45 PM, Rob Sherwood wrote:
Hi Larry,

WSJT X does not have multiple tones transmitted at the same time. You can run a class C amplifier with JT9, for example, a mode I use on 630 meters. There may well be other modes that use multiple simultaneous tones, in which case linearity would matter. Hopefully others will chime in who have more knowledge of all things digital.

Rob, NC0B

Dave Cole

Hi Rob,

I have also seen phantom signals from F/H stations, I have not decoded them however...

73, and thanks,
Dave (NK7Z)
https://www.nk7z.net
ARRL Volunteer Examiner
ARRL Technical Specialist
ARRL Asst. Director, NW Division, Technical Resources

On 11/8/19 7:25 AM, Rob Sherwood wrote:
Hi Dave,
That is a good point, if there are multiple simultaneous FT8 and FT4 (?) QSOs going on, then yes there would be intermodulation products. Considering the IMD products are only down 20 some dB, that would produce a lot of phantom signals that could decode on the other end. I definitely decode line-related phantom signals on 630m JT9 on strong signals. I have decoded as many as 2 phantom signals on each side of the mail signal.
Rob, NC0B
-----Original Message-----
From: Elecraft-K3@groups.io [mailto:Elecraft-K3@groups.io] On Behalf Of Dave Cole
Sent: Friday, November 08, 2019 4:44 AM
To: Elecraft-K3@groups.io
Subject: Re: [Elecraft-K3] Power supply adjustments
Hi Rob,
Does not WSJT-X transmit multiple tones in Fox/Hound mode?
73, and thanks,
Dave (NK7Z)
https://www.nk7z.net
ARRL Volunteer Examiner
ARRL Technical Specialist
ARRL Asst. Director, NW Division, Technical Resources
On 11/7/19 6:45 PM, Rob Sherwood wrote:
Hi Larry,

WSJT X does not have multiple tones transmitted at the same time. You can run a class C amplifier with JT9, for example, a mode I use on 630 meters. There may well be other modes that use multiple simultaneous tones, in which case linearity would matter. Hopefully others will chime in who have more knowledge of all things digital.

Rob, NC0B