Topics

QUESTION for All - EASY BITX - HP 2800 series Schottky Diodes for Mixer


 

I would like to know if I can use HP 2800 series Schottky diodes in the EASY BITX mixers circuits.
Can anyone give me answers?

73 de --
IW4AJR Loris


Rafael Pinto [PU1OWL]
 

Loris,

By looking at the datasheet, you can use it. The thing is you will have to match the diodes Vf very carefully. Using matched diodes like the BAT41 halfs your work!  

73

Rafael - PU1OWL


 

Hi Rafael,

It will not be easy to combine four diodes, but I have to buy the BAT41s, while I have the HP2835s in the drawer (at least forty) ... I hope to be able to find two groups of 4 diodes at least very similar.
The HP gave a margin of plus or minus 20mV of Vf between two diodes in match pair, is this an acceptable term? that is, what can be the maximum deviation between two diodes for a match pair?

On Tue, Jan 12, 2021 at 12:37 PM, <rafaelgcpp@...> wrote:
Loris,

By looking at the datasheet, you can use it. The thing is you will have to match the diodes Vf very carefully. Using matched diodes like the BAT41 halfs your work!  

73

Rafael - PU1OWL

--
IW4AJR Loris


Cristian Radulescu
 

Hi all,
Sorry to barge in like this.
12-13 years ago when I assembled my first, ugly style, BitX, I remember experimenting a lot with these mixers.
What I found is, one can not match the matched mixer rings that exist, produced by various companies.
What I ended up using was a diode ring IC "extracted" from a Motorola CP040 or GP300 VHF handy placed under a small shield close to the receiver part, between two trifilar transformers (obviously a mixer configuration), 8 pin IC chip produced by NEC or JRC.
I am sure one can recover boards from these old handies for parts recovery.
I had excellent results with it.
73's de YO3IAZ, Cristian

On Tuesday, January 12, 2021, 02:18:30 PM GMT+2, Loris IW4AJR via groups.io <lorisbollina@...> wrote:


Hi Rafael,

It will not be easy to combine four diodes, but I have to buy the BAT41s, while I have the HP2835s in the drawer (at least forty) ... I hope to be able to find two groups of 4 diodes at least very similar.
The HP gave a margin of plus or minus 20mV of Vf between two diodes in match pair, is this an acceptable term? that is, what can be the maximum deviation between two diodes for a match pair?

On Tue, Jan 12, 2021 at 12:37 PM, <rafaelgcpp@...> wrote:
Loris,

By looking at the datasheet, you can use it. The thing is you will have to match the diodes Vf very carefully. Using matched diodes like the BAT41 halfs your work!  

73

Rafael - PU1OWL

--
IW4AJR Loris


 

Hi Cristian,
I know very well these mixer modules, there are many brands and in different solutions (SMD, with through holes, with SMA connectors, etc.) but the aim is to use low-cost materials (actually at no cost, since I have them in the drawer) without modifying the EASY-BITX PCB which is fine as is.
For this purpose I am attaching some files with lists of various diodes also containing some application notes and some diode data.
Thanks anyway for the answer, but the way of the prefab mixers has been known to me for a long time.
73 de
--
IW4AJR Loris


Arv Evans
 

Over the years I have used my own matched diodes in mixers.  Matching is
usually at a fixed current with a high impedance VOM measuring forward
voltage drop.  On a few circuits I generated forward and reverse current (leakage)
curves at DC, and at RF but did not see much gained by excessive measurement.
Key for me has been matching the "knee point" where the diodes go into conduction, and matching the leakage curve if the diodes will be subject to high RF levels.

Arv
_._


On Tue, Jan 12, 2021 at 3:56 AM Loris IW4AJR via groups.io <lorisbollina=mac.com@groups.io> wrote:
I would like to know if I can use HP 2800 series Schottky diodes in the EASY BITX mixers circuits.
Can anyone give me answers?

73 de --
IW4AJR Loris


Bob Lunsford
 

It is always better to use diodes and transistors made on the same substrate.

I remember replacing the PA transistor on an HW-8 with a dual transistor placed in parallel, doubled the power capability and since the  two transistors were on the same substrate, they were perfectly matched. If an IC can be found with four diodes, then finding matched diodes will not be required, if done on the same substrate ... IF such a thing exists.

Bob — KK5R

On Tuesday, January 12, 2021, 8:09:17 AM EST, Cristian Radulescu via groups.io <criradul2001@...> wrote:


Hi all,
Sorry to barge in like this.
12-13 years ago when I assembled my first, ugly style, BitX, I remember experimenting a lot with these mixers.
What I found is, one can not match the matched mixer rings that exist, produced by various companies.
What I ended up using was a diode ring IC "extracted" from a Motorola CP040 or GP300 VHF handy placed under a small shield close to the receiver part, between two trifilar transformers (obviously a mixer configuration), 8 pin IC chip produced by NEC or JRC.
I am sure one can recover boards from these old handies for parts recovery.
I had excellent results with it.
73's de YO3IAZ, Cristian

On Tuesday, January 12, 2021, 02:18:30 PM GMT+2, Loris IW4AJR via groups.io <lorisbollina@...> wrote:


Hi Rafael,

It will not be easy to combine four diodes, but I have to buy the BAT41s, while I have the HP2835s in the drawer (at least forty) ... I hope to be able to find two groups of 4 diodes at least very similar.
The HP gave a margin of plus or minus 20mV of Vf between two diodes in match pair, is this an acceptable term? that is, what can be the maximum deviation between two diodes for a match pair?

On Tue, Jan 12, 2021 at 12:37 PM, <rafaelgcpp@...> wrote:
Loris,

By looking at the datasheet, you can use it. The thing is you will have to match the diodes Vf very carefully. Using matched diodes like the BAT41 halfs your work!  

73

Rafael - PU1OWL

--
IW4AJR Loris


Gordon Gibby
 

In the old days, I think we used a potentiometer and maybe a variable capacitor to balance out these mixers. I’ve got a Heathkit SB 102 that I soldered together as a high school kid, and it has four individual diodes and I remember nulling out the carrier

Surely it can still be successful with all of the advanced semiconductors we have now. 




On Jan 12, 2021, at 16:29, Bob Lunsford via groups.io <nocrud222@...> wrote:


It is always better to use diodes and transistors made on the same substrate.

I remember replacing the PA transistor on an HW-8 with a dual transistor placed in parallel, doubled the power capability and since the  two transistors were on the same substrate, they were perfectly matched. If an IC can be found with four diodes, then finding matched diodes will not be required, if done on the same substrate ... IF such a thing exists.

Bob — KK5R

On Tuesday, January 12, 2021, 8:09:17 AM EST, Cristian Radulescu via groups.io <criradul2001@...> wrote:


Hi all,
Sorry to barge in like this.
12-13 years ago when I assembled my first, ugly style, BitX, I remember experimenting a lot with these mixers.
What I found is, one can not match the matched mixer rings that exist, produced by various companies.
What I ended up using was a diode ring IC "extracted" from a Motorola CP040 or GP300 VHF handy placed under a small shield close to the receiver part, between two trifilar transformers (obviously a mixer configuration), 8 pin IC chip produced by NEC or JRC.
I am sure one can recover boards from these old handies for parts recovery.
I had excellent results with it.
73's de YO3IAZ, Cristian

On Tuesday, January 12, 2021, 02:18:30 PM GMT+2, Loris IW4AJR via groups.io <lorisbollina@...> wrote:


Hi Rafael,

It will not be easy to combine four diodes, but I have to buy the BAT41s, while I have the HP2835s in the drawer (at least forty) ... I hope to be able to find two groups of 4 diodes at least very similar.
The HP gave a margin of plus or minus 20mV of Vf between two diodes in match pair, is this an acceptable term? that is, what can be the maximum deviation between two diodes for a match pair?

On Tue, Jan 12, 2021 at 12:37 PM, <rafaelgcpp@...> wrote:
Loris,

By looking at the datasheet, you can use it. The thing is you will have to match the diodes Vf very carefully. Using matched diodes like the BAT41 halfs your work!  

73

Rafael - PU1OWL

--
IW4AJR Loris


Arv Evans
 

Gordon

Same here.  Those integrated diode IC's are a big improvement but even
ones from the same corner of the same die are probably not perfectly
matched after they have been manually soldered onto a PCB.  Even for
those it seems better to have a balancing network to insure we get the
best balance after everything is assembled.  Maybe this is just a feel-good
situation but part of optimizing a build is feeling good about how it went
together and how it was adjusted after the build.  This seems especially
true if the mixer is a 2-diode version instead of the full ring of 4 diodes.

Arv
_._


On Tue, Jan 12, 2021 at 2:40 PM Gordon Gibby <docvacuumtubes@...> wrote:
In the old days, I think we used a potentiometer and maybe a variable capacitor to balance out these mixers. I’ve got a Heathkit SB 102 that I soldered together as a high school kid, and it has four individual diodes and I remember nulling out the carrier

Surely it can still be successful with all of the advanced semiconductors we have now. 




On Jan 12, 2021, at 16:29, Bob Lunsford via groups.io <nocrud222=yahoo.com@groups.io> wrote:


It is always better to use diodes and transistors made on the same substrate.

I remember replacing the PA transistor on an HW-8 with a dual transistor placed in parallel, doubled the power capability and since the  two transistors were on the same substrate, they were perfectly matched. If an IC can be found with four diodes, then finding matched diodes will not be required, if done on the same substrate ... IF such a thing exists.

Bob — KK5R

On Tuesday, January 12, 2021, 8:09:17 AM EST, Cristian Radulescu via groups.io <criradul2001=yahoo.com@groups.io> wrote:


Hi all,
Sorry to barge in like this.
12-13 years ago when I assembled my first, ugly style, BitX, I remember experimenting a lot with these mixers.
What I found is, one can not match the matched mixer rings that exist, produced by various companies.
What I ended up using was a diode ring IC "extracted" from a Motorola CP040 or GP300 VHF handy placed under a small shield close to the receiver part, between two trifilar transformers (obviously a mixer configuration), 8 pin IC chip produced by NEC or JRC.
I am sure one can recover boards from these old handies for parts recovery.
I had excellent results with it.
73's de YO3IAZ, Cristian

On Tuesday, January 12, 2021, 02:18:30 PM GMT+2, Loris IW4AJR via groups.io <lorisbollina=mac.com@groups.io> wrote:


Hi Rafael,

It will not be easy to combine four diodes, but I have to buy the BAT41s, while I have the HP2835s in the drawer (at least forty) ... I hope to be able to find two groups of 4 diodes at least very similar.
The HP gave a margin of plus or minus 20mV of Vf between two diodes in match pair, is this an acceptable term? that is, what can be the maximum deviation between two diodes for a match pair?

On Tue, Jan 12, 2021 at 12:37 PM, <rafaelgcpp@...> wrote:
Loris,

By looking at the datasheet, you can use it. The thing is you will have to match the diodes Vf very carefully. Using matched diodes like the BAT41 halfs your work!  

73

Rafael - PU1OWL

--
IW4AJR Loris


 

Hi everyone ...
Thanks for all the answers!
But at this point I am wondering :
why not use an integrated active mixer? 
such as, for example, a NE612 or similar!
I understand it's not bidirectional, but two NE612 ICs really cost more than two toroidal transformers and a diode bridge?
It seems to me that they would solve many problems of balance and of component macht.
Have anyone information of malfunctions of these active mixers?
Has anyone ever tried to use them in bidirectional medium frequency circuits?
Thanks to all ... de
--
IW4AJR Loris


Henning Weddig
 

Am 13.01.2021 um 10:11 schrieb Loris IW4AJR via groups.io:
Hi everyone ...
Thanks for all the answers!
But at this point I am wondering :
why not use an integrated active mixer?
such as, for example, a NE612 or similar!
I understand it's not bidirectional, but two NE612 ICs really cost more than two toroidal transformers and a diode bridge?
It seems to me that they would solve many problems of balance and of component macht.
Have anyone information of malfunctions of these active mixers?
Has anyone ever tried to use them in bidirectional medium frequency circuits?
Thanks to all ... de
--
IW4AJR Loris
the NE612 or other gilbert cell mixers suffer from a too low IP3.


 

Hi Henning,
Looking around in the various datasheets it seems to me that the IP3 story is not very satisfying.
In the HF, even the professional balanced mixers of the Minicircuits and MaCom give an IP3 ranging from -15 to a maximum of -20 dB, a simple NE612 still guarantees an IP3 of -13dB, I don't think that a manual construction of toroidal transformers and a ring of 4 diodes (more or less selected) can guarantee much more than the IP3 of a NE612.
Is it not by chance that all the experimenters have been very little interested in how to interface the NE612 both from the point of view of the input and output impedances and the power level of the LO?
I believe that a more accurate analysis would lead to very different results! ... is't it?
73 de
--
IW4AJR Loris


iz oos
 

According to Franco Rota rfmicrowave.it 
It can be used as detector for applications up to 2 GHz, as mixer up to 1.5 GHz, often used also as RX input protection at HF and VHF thanks to its low capacity and high switching speed. 
Indeed I had used them for RF protection. 

Regarding the SA602 of you consider the gain Is not much forse than usual SBL or similar mixers.

Il mer 13 gen 2021 12:06 PM Loris IW4AJR via groups.io <lorisbollina=mac.com@groups.io> ha scritto:
Hi Henning,
Looking around in the various datasheets it seems to me that the IP3 story is not very satisfying.
In the HF, even the professional balanced mixers of the Minicircuits and MaCom give an IP3 ranging from -15 to a maximum of -20 dB, a simple NE612 still guarantees an IP3 of -13dB, I don't think that a manual construction of toroidal transformers and a ring of 4 diodes (more or less selected) can guarantee much more than the IP3 of a NE612.
Is it not by chance that all the experimenters have been very little interested in how to interface the NE612 both from the point of view of the input and output impedances and the power level of the LO?
I believe that a more accurate analysis would lead to very different results! ... is't it?
73 de
--
IW4AJR Loris


 

Yes for sure,

Here is a good article to follow for matching them:  http://www.seekic.com/circuit_diagram/Measuring_and_Test_Circuit/Diode_Matching_for_Mixers.html

Make sure to handle the diodes with tweezers as body heat from your fingers will affect the measurement a lot.

73,   Mark


Jerry Gaffke
 

Loris,

The very cheap ADE-1 level 7 diode ring mixer has an IP3 of around 15dBm (not dB),
good through 500mhz.  Not sure why the uBitx mixers could not approach 15dbm at 3-30mhz.

The SA612 draws a minuscule  2.4ma from a 5v supply, requires less LO drive,
and gives 17dB of signal gain (the ADE-1 incurs a 5dB loss).
But those advantages come with compromises, the dynamic range of the SA612 is limited.

Jerry, KE7ER


On Wed, Jan 13, 2021 at 03:04 AM, Loris IW4AJR wrote:
Hi Henning,
Looking around in the various datasheets it seems to me that the IP3 story is not very satisfying.
In the HF, even the professional balanced mixers of the Minicircuits and MaCom give an IP3 ranging from -15 to a maximum of -20 dB, a simple NE612 still guarantees an IP3 of -13dB, I don't think that a manual construction of toroidal transformers and a ring of 4 diodes (more or less selected) can guarantee much more than the IP3 of a NE612.
Is it not by chance that all the experimenters have been very little interested in how to interface the NE612 both from the point of view of the input and output impedances and the power level of the LO?
I believe that a more accurate analysis would lead to very different results! ... is't it?
73 de


Curt
 

Very interesting topic, and I don't how much recent work has been analyzed by amateurs since the texts by w7zoi. A difference of a few dBm in third order intercept isn't a huge consequence for most hf amateur operation, but 6 or 10 dB difference might be noted. I am thinking the matching of diodes is mostly for balance in the network, and this increases isolation among the ports. Maybe this is of consequence in upper hf for ubitx to address the well known leakage spurs on ssb transmit, it may not he as critical at low hf. Without analysis I don't know how important diode matching may be for ubitx receive.

Always see if TOI is referenced to input or output, as there is a meaningful difference in performance of sa612 style versus diode rings. However often my pfr3 using these mixers does well enough in uncongested bands. The sw40 does better putting the only gain control in front of the sa612. In my judgment the ubitx linearity is not vastly inferior to my k2, with its main enhancement being a higher linearity IF amp right after the mixer.

Not many home constructed rigs use mosfet mixers, but these can be higher performing. The cyclone that was kitted by 4stateqrp in US uses a pair of large mosfets for its first mixer, in fact they are also the transmit PA.

For a single band bitx at low hf I am thinking precise balance might not be that critical, so try what you have. They may still be nice diodes for higher bands.

73 curt wb8yyy


Ashhar Farhan
 

Loris,
The minicircuit mixers have an IIP3 (input intercept) of +15dbm. I am not sure where you read it as -20db  (it should be dbm). the NE612 quotes it at -13 dbm but the measured results are closer to -15 dbm. The dynamic range difference is very dramatic. The NE612 loads very easily. 
It still has its uses. I am using it to build a new transponder where the current consumption is an issue and very loud signals are not expected (in a satellite, the nearest station is 600 km away). It is also useful on VHF and UHF where stations run just a few watts each and gain is an advantage. That apart, using the NE612 in front-ends is not a good idea at all. It can saturate even with atmospheric noise. Four generic diodes in a double balanced mixer followed by a 2N3904 feedback amplifier can give far far better dynamic range for a lower cost and similar noise figure.
- f


On Wed, Jan 13, 2021 at 4:36 PM Loris IW4AJR via groups.io <lorisbollina=mac.com@groups.io> wrote:
Hi Henning,
Looking around in the various datasheets it seems to me that the IP3 story is not very satisfying.
In the HF, even the professional balanced mixers of the Minicircuits and MaCom give an IP3 ranging from -15 to a maximum of -20 dB, a simple NE612 still guarantees an IP3 of -13dB, I don't think that a manual construction of toroidal transformers and a ring of 4 diodes (more or less selected) can guarantee much more than the IP3 of a NE612.
Is it not by chance that all the experimenters have been very little interested in how to interface the NE612 both from the point of view of the input and output impedances and the power level of the LO?
I believe that a more accurate analysis would lead to very different results! ... is't it?
73 de
--
IW4AJR Loris


Ashhar Farhan
 

the ubitx mixers report 15 dbm of iip3, it is the post mixer amplifier that lets it down. The output intercept of the termination insensitive mixer +20 dbm. Hence, the input intercept will be around +5dbm (minus the gain). The combined intercept will be lowered. If we used a better amplifier (costing us the effort to wind another toroid  as the output balun) with an inductive load instead of a resistor, we could have managed with far better intercept. I had to balance ease of duplication and performance. 
Not that anyone has complained about the lack of better dynamic range. but we are coasting around the performance of rigs in the mid 1990s. On my bench is a new rig coming up that is built for high performance, hybrid sdr capabilities. The measured intercept on this one is +25 dbm. Very solid.

- f

On Thu, Jan 14, 2021 at 11:54 AM Ashhar Farhan <farhanbox@...> wrote:
Loris,
The minicircuit mixers have an IIP3 (input intercept) of +15dbm. I am not sure where you read it as -20db  (it should be dbm). the NE612 quotes it at -13 dbm but the measured results are closer to -15 dbm. The dynamic range difference is very dramatic. The NE612 loads very easily. 
It still has its uses. I am using it to build a new transponder where the current consumption is an issue and very loud signals are not expected (in a satellite, the nearest station is 600 km away). It is also useful on VHF and UHF where stations run just a few watts each and gain is an advantage. That apart, using the NE612 in front-ends is not a good idea at all. It can saturate even with atmospheric noise. Four generic diodes in a double balanced mixer followed by a 2N3904 feedback amplifier can give far far better dynamic range for a lower cost and similar noise figure.
- f

On Wed, Jan 13, 2021 at 4:36 PM Loris IW4AJR via groups.io <lorisbollina=mac.com@groups.io> wrote:
Hi Henning,
Looking around in the various datasheets it seems to me that the IP3 story is not very satisfying.
In the HF, even the professional balanced mixers of the Minicircuits and MaCom give an IP3 ranging from -15 to a maximum of -20 dB, a simple NE612 still guarantees an IP3 of -13dB, I don't think that a manual construction of toroidal transformers and a ring of 4 diodes (more or less selected) can guarantee much more than the IP3 of a NE612.
Is it not by chance that all the experimenters have been very little interested in how to interface the NE612 both from the point of view of the input and output impedances and the power level of the LO?
I believe that a more accurate analysis would lead to very different results! ... is't it?
73 de
--
IW4AJR Loris


 

Hi Ashhar
 
you probably have not read the chat title correctly, I am not referring to the µBITX of your production, but to the EASY BITX (40m) that I am mounting, your mixers are fine as they are, in the EASY version the inputs and outputs are instead , toroidal transformers, are wound and twisted by hand, this is my concern! ... I don't think it's possible to get high IP3 performance from hand-wound toorids (by a hobbyist) and a 1N4148 quad diode.
So I was looking for information on possible improvements of these mixers! ... Can you tell me where I can get the pre-wound toroid you use in the µBITX-V6 card?
By the way, the µBITX-V6 that I have already assembled is fine as it is, in fact I have decided to also remove the AGG (which I have tried but it works very badly) and to use µBITX-V6 as you have built it, any change only risks getting worse instead of bringing some benefit.
Thanks for the reply 73 de
--
IW4AJR Loris


Jerry Gaffke
 

HFSignals has never sold individual parts.
The cost to ship from India makes that a bad idea.

There is no special trick to winding those torroids.
I doubt that how tightly they are twisted or how tightly they are wound on the torroid
makes much difference at all.

If you are worried about mixer performance, I'd recommend substituting an ADE-1 or SBL-1
level 7 diode ring mixer, see if that makes any difference.
If you buy a genuine part from Minicircuits, the performance to expect is documented.
It would be interesting to compare results between a hand built mixer such as found on
the uBitx and a manufactured mixer.  I'd guess that the primary difference we would see
in the frequency range we are interested in would be in the rejection of undesired products
due to not having all four diodes on the same die for best matching.

Jerry, KE7ER


On Thu, Jan 14, 2021 at 06:05 AM, Loris IW4AJR wrote:
Hi Ashhar
 
you probably have not read the chat title correctly, I am not referring to the µBITX of your production, but to the EASY BITX (40m) that I am mounting, your mixers are fine as they are, in the EASY version the inputs and outputs are instead , toroidal transformers, are wound and twisted by hand, this is my concern! ... I don't think it's possible to get high IP3 performance from hand-wound toorids (by a hobbyist) and a 1N4148 quad diode.
So I was looking for information on possible improvements of these mixers! ... Can you tell me where I can get the pre-wound toroid you use in the µBITX-V6 card?
By the way, the µBITX-V6 that I have already assembled is fine as it is, in fact I have decided to also remove the AGG (which I have tried but it works very badly) and to use µBITX-V6 as you have built it, any change only risks getting worse instead of bringing some benefit.
Thanks for the reply 73 de
--
IW4AJR Loris