Snubber vs safety cap in 7854.


Martin Hodge
 

My 7854 recently started to act up and after investigation I found the +15v regulated supply was sagging to 14.2v and had some ~50kHz 20mV ripple. Probing deeper into the SMPS I further found the unreg +17v supplying the +15v regulator had the same 50kHz ripple but at 168mV! The problem turned out to be C142 on the A12 control rectifier board. This 330uF Sprague axial lead capacitor had dried up and was not doing it's job. I the process of replacing most of the capacitors in the SMPS I've come across an infamous Rifa safety cap across the mains input. After ordering a modern X2 rated replacement for it I realized the Rifa is not a safety cap, but a snubber. It has a built-in 22 ohm resistor in series with the .1uF capacitor. Looking at the schematic, I see Tek built a snubber out of discrete parts, but at some later point replaced the parts with the Rifa.

The question... how important is the resistor? Will the new X2 safety cap by itself be okay, or should I endeavor to add the series resistance?


Jean-Paul
 

Please include a screen shot of the schematic and of the capacitor,

Jon


Dave Daniel
 

I would use an exact (or very close) replacement. So-called "RIFA" caps have a specific reactance curve designed for that particular use; they are designed to be open-circuit at mains frequencies and nearly a short-circuit at high frequencies, with the reactance decreasing as the frequency increases. Their purpose is to provide some degree of filtering by shorting out  higher frequencies, i.e., noise.

DaveD

On 4/10/2021 9:34 PM, Martin Hodge wrote:
My 7854 recently started to act up and after investigation I found the +15v regulated supply was sagging to 14.2v and had some ~50kHz 20mV ripple. Probing deeper into the SMPS I further found the unreg +17v supplying the +15v regulator had the same 50kHz ripple but at 168mV! The problem turned out to be C142 on the A12 control rectifier board. This 330uF Sprague axial lead capacitor had dried up and was not doing it's job. I the process of replacing most of the capacitors in the SMPS I've come across an infamous Rifa safety cap across the mains input. After ordering a modern X2 rated replacement for it I realized the Rifa is not a safety cap, but a snubber. It has a built-in 22 ohm resistor in series with the .1uF capacitor. Looking at the schematic, I see Tek built a snubber out of discrete parts, but at some later point replaced the parts with the Rifa.

The question... how important is the resistor? Will the new X2 safety cap by itself be okay, or should I endeavor to add the series resistance?



--
This email has been checked for viruses by Avast antivirus software.
https://www.avast.com/antivirus


 

On Sun, Apr 11, 2021 at 09:39 AM, Dave Daniel wrote:
with the reactance decreasing as the frequency increases.
Well, yes, that's how capacitors work :)

Or did you mean, self-resonant at a particular frequency? Above that frequency the cap would appear inductive, though, so I'm not sure that would help.


Dave Daniel
 

A *particular* reactance curve designed for that particular application. I wasn't writing in generalities.

I did not write "self-resonant" because that is not the characteristic of RIFA caps that is important *for this application*.

DaveD

On 4/11/2021 11:07 AM, Charles wrote:
On Sun, Apr 11, 2021 at 09:39 AM, Dave Daniel wrote:
with the reactance decreasing as the frequency increases.
Well, yes, that's how capacitors work :)

Or did you mean, self-resonant at a particular frequency? Above that frequency the cap would appear inductive, though, so I'm not sure that would help.



--
This email has been checked for viruses by Avast antivirus software.
https://www.avast.com/antivirus


Jean-Paul
 

Hello Martin:

An "X" rated cap placed across the mains is NOT a snubber. Snubbers are inside an SMPS to reduce the inductive transients from switching an inductive load.

Instead they are specific to EMI and transient filters: Function: To terminate the EMI filter, Normal Mode, to reduce any incomning tranisents. The caps you mention had the known fault of fires and explosions after long use due to plastic crazing.

The series R was added specific to that RIFA cap to act as a fuse and reduce the risks.

Any PROPERLY RATED modern X cap without any series R is fine.

We use 280VAC rated 100 or 470 nF, Panasonic, and many others make these.

Kind Regards,

Jon


Martin Hodge
 

See these photos and schematic here
https://groups.io/g/TekScopes/album?id=262875

R5, C5, and C6 are the parts in question.
R5 = 27 ohm 1-watt
C5,C6 = .02 uf 1400 VDC

The Rifa was soldered in where R5 would be and a jumper is where C5 would be. What I will do is place the .1 uF X2 cap where the Rifa was, remove the jumper and parallel two 47 ohm 1/2 watt resistors where C5 and C6 would be. Seeing how the Rifa Tek chose doesn't exactly match the values in the schematic it doesn't appear to be too critical.


Jean-Paul
 

Martin I would use flame proof resistors, metal oxide, perhaps not carbon or metal film.

All the usual disty have them, DK, Mouser, etc. Space the resistors about 1 cm off the board just in case!

There is a safety issue especially on 240V mains.

Kind Regards,

Jon


Stephen Bell
 

I had the same problem in my 7854 (S/No. B042065) but worse. One of the Sprague 330uF capacitors had dried up so much that the capacitance had reduced to 0.5uF and the ESR had increased to 15kohm. The Sprague capacitors used by Tek were inadequate for the circuit ripple current: I measured 1.3ARMS ripple current whereas the Sprague capacitor would probably have only been good for half that current.

In the manuals I have the BOM lists a Panasonic part number for these capacitors rather than the Sprague capacitors that were fitted to my 7854. I presume the inadequacy in the Sprague capacitors was recognised early on and the part was upgraded for latter production although there is no mention in the BOM of an earlier versions using the Sprague part numbers. It would be interesting to know at what point the capacitors were upgraded and whether the same problem affects the 7104 which uses essentially the same power supply as the 7854.


Tom Lee
 

As others have said, that's the textbook reactance behavior of capacitors, so that's not the key feature.

The characteristic that distinguishes safety-rated X and Y capacitors from ordinary caps isn't the reactance behavior, it's the way that they are designed to fail. This is an important consideration if the cap is connected to the line. An X cap is supposed to fail shorted, and a Y cap is supposed to fail open. Depending on whether the cap is connected between hot and neutral, or between hot and ground, one or the other behavior is better for safety.

Tom

Sent from an iThing, so please forgive the typos and brevity

On Apr 11, 2021, at 7:39, "Dave Daniel" <kc0wjn@gmail.com> wrote:

I would use an exact (or very close) replacement. So-called "RIFA" caps have a specific reactance curve designed for that particular use; they are designed to be open-circuit at mains frequencies and nearly a short-circuit at high frequencies, with the reactance decreasing as the frequency increases. Their purpose is to provide some degree of filtering by shorting out higher frequencies, i.e., noise.

DaveD


On 4/10/2021 9:34 PM, Martin Hodge wrote:
My 7854 recently started to act up and after investigation I found the +15v regulated supply was sagging to 14.2v and had some ~50kHz 20mV ripple. Probing deeper into the SMPS I further found the unreg +17v supplying the +15v regulator had the same 50kHz ripple but at 168mV! The problem turned out to be C142 on the A12 control rectifier board. This 330uF Sprague axial lead capacitor had dried up and was not doing it's job. I the process of replacing most of the capacitors in the SMPS I've come across an infamous Rifa safety cap across the mains input. After ordering a modern X2 rated replacement for it I realized the Rifa is not a safety cap, but a snubber. It has a built-in 22 ohm resistor in series with the .1uF capacitor. Looking at the schematic, I see Tek built a snubber out of discrete parts, but at some later point replaced the parts with the Rifa.

The question... how important is the resistor? Will the new X2 safety cap by itself be okay, or should I endeavor to add the series resistance?




--
This email has been checked for viruses by Avast antivirus software.
https://www.avast.com/antivirus






 

And a Y cap can be safely used to replace an X cap.
It's just a matter of cost - Y caps are ALWAYS more expensive than X caps.

But an X cap may NEVER be used to replace a Y cap!


Roy Thistle
 

On Sun, Apr 11, 2021 at 08:21 AM, Jean-Paul wrote:


An "X" rated cap placed across the mains is NOT a snubber.
Hi Jean-Paul+Martin:
Well looking at the circuit... it seems to me... that R5,C5,C6 do constitute a "snubber." and the Rifa cap was added as a bodge. Maybe that Rifa bodge (or the planned bodge) is good enough; but, the bridge (CR15) might eventually go.
--
Roy Thistle


Martin Hodge
 

Thanks for all the help, I've elected to install the cap without resistor. I feel safe enough considering it's only connected to mains when the unit is on, it's behind two layers of aluminum chassis, and I'm on 120v mains.

Roy, what do you mean the bridge might eventually go?

(Terse reply because of horrible virtual keyboard)


Dan G
 

I've elected to install the cap without resistor. I
feel safe enough considering it's only connected to mains when the unit is on,
it's behind two layers of aluminum chassis, and I'm on 120v mains.
I suspect the purpose of the series resistor is not just for safety in case
C5 shorts out, but also to limit the dv/dt across the capacitor.

Tektronix spec'd this R5 limiter both in the original discrete design, and in the
post-B070000 change via the combined RC network. Simply ignoring this,
and replacing the RC network with a capacitor and no current limiter seems...
not ideal.


dan


Dan G
 

On Sun, Apr 11, 2021 at 11:48 AM, Martin Hodge wrote:

See these photos and schematic here
https://groups.io/g/TekScopes/album?id=262875
I have added an updated version of the schematic to your album.
It shows the relevant change that Tektronix made at serial number B070000.

The electrical parts list identifies the new A23C5 as:

Tek p/n: 119-1168-00
Description: Capacitor-Res: 0.1uF,20% & 22 Ohm,10%,250V
Manufacturer: Electro Cube Inc.
Mfr p/n: RG1782-1


dan


Martin Hodge
 

On Tue, Apr 13, 2021 at 06:08 PM, Dan G wrote:


On Sun, Apr 11, 2021 at 11:48 AM, Martin Hodge wrote:

See these photos and schematic here
https://groups.io/g/TekScopes/album?id=262875
I have added an updated version of the schematic to your album.
It shows the relevant change that Tektronix made at serial number B070000.

The electrical parts list identifies the new A23C5 as:

Tek p/n: 119-1168-00
Description: Capacitor-Res: 0.1uF,20% & 22 Ohm,10%,250V
Manufacturer: Electro Cube Inc.
Mfr p/n: RG1782-1


dan
Where did you find that schematic? My SN is B074375


Dan G
 

On Tue, Apr 13, 2021 at 10:19 PM, Martin Hodge wrote:

Where did you find that schematic?
On my bookshelf :)

The image I uploaded is a scan of my (original) paper manual.


Roy Thistle
 

Hi Martin:
Well there is already a line filter before the switch ... but, I don't know if the line filter has line to line X... and line to chassis Y... you could check.
The only reason for R5, C5, C6 ... that I can see is to protect the bridge rectifier.
Are you saying that Tektronix moded those out in an updated version of the schematic?

--
Roy Thistle


Martin Hodge
 

On Thu, Apr 15, 2021 at 06:04 PM, Roy Thistle wrote:


Hi Martin:
Well there is already a line filter before the switch ... but, I don't know if
the line filter has line to line X... and line to chassis Y... you could
check.
The only reason for R5, C5, C6 ... that I can see is to protect the bridge
rectifier.
Are you saying that Tektronix moded those out in an updated version of the
schematic?

--
Roy Thistle
Well hey, it's right there in the manual:

"LINE INPUT
Power is applied through line filter FL10, line fuse F10,
and POWER switch S10. The line filter is designed to
keep powerline interference from entering the
instrument and to keep the approximate 25-kilohertz
Inverter signal from entering the power line. Components
R5, C5 and C6 suppress reverse-recovery transients of
CR15"

I'll turn 50 this year. You'd think by now I'd have learned to read the manual.
So, yeah, Tek replaced those three components with one monolithic package made by Rifa.