Sherwood S-5000 restoration

bberkom

AK Subscriber
Subscriber
As a part of a purchase of some tube gear, I picked up this really nice Sherwood S-5000 integrated. The lady actually gave it to me for free because I bought a bunch of other stuff and got a friend to come buy the rest of her late husband's hoard. It is completely original and fully functional after bringing it up and doing some quick testing. No hum at all, and it really sounds great. All 5 12ax7's are the original telefunkens and test strong. 7199's are original, as is the rectifier. I can barely make out the letters "SHERW" on the rectifier. The outputs are not original, but they are nice Sovtek EL84M's that must be close to brand new.

For restoration, I plan to add a CL80, replace the electrolytics and the rectifier, and see how it performs. I am going to use new caps under the chassis for the small 30/20 can and the common-positive bias can. Since there is so little room in the chassis, I ordered a CE can cap from AES for the main power supply can. I also ordered new coupling caps that I may replace preventatively.

Most of that seems pretty straightforward, but I already have a question about replacing the rectifier. I was planning to use two 1.5a 1000v diodes to create a new bridge on a terminal strip for the rectifier replacement as see on this thread, but I have some GBU4G 4 amp rectifiers that I have used on a fisher and a scott. These have a DC+, DC-, and two AC ~ legs coming off of it, but the original rectifier has just the two ~ leads from the transformer and one lead that goes to the bias supply. Is the original rectifier grounded through its case to the chassis and would I just have to ground one of the DC legs of the GBU4G rectifier if I use that instead of the diodes? I may have to use the diodes, just thought using the GBU4G may be cleaner and easier if possible.

Since the voltage from the new rectifier will be higher than the old rectifier, I may also need help placing a voltage dropping resistor to get the voltage down. I ordered a handful of different 5W resistors, I think around 8-10ohm to try. Will a 5W resistor be suitable for this?

Guess I have more questions than I thought. I should have parts by tomorrow and start on it this weekend. Thanks for any help.
 

Attachments

  • IMG_6378.jpg
    IMG_6378.jpg
    57 KB · Views: 282
  • IMG_6379.jpg
    IMG_6379.jpg
    35.9 KB · Views: 264
  • IMG_6380.jpg
    IMG_6380.jpg
    79.1 KB · Views: 278
  • IMG_6278.JPG
    IMG_6278.JPG
    84 KB · Views: 276
It sounds to me like the GBU4G is a full wave bridge rectifier while what was used by Sherwood for the bias supply was a half-wave, which is common for bias supplies. In fact many just use one diode coming off one leg of the HT winding.

You could use the GBU4G but no, you wouldn't ground either end. You'd leave the "+" end floating (heat shrink the lead) and connect the "-" end to your bias supply to get negative DC voltage there. Just like using two individual diodes, lol, and half of the GBU4G is unused. You could save them for the next German table radio you restore.

Anyway, having more negative voltage available for a bias supply is better than having less, as long as you are able to make it adjustable and have a way to measure cathode current in your power tubes. More negative voltage than required means the tubes will, initially, be biased colder. This is a better starting point that when they are biased hotter. Then you experiment with the resistors-to-ground in the bias supply and the adjustment pots to get them in a range that you like. I like to temporarily replace the resistor to ground with a linear pot, and set it so it where I get the most usable range out of the main adjustment pot. Then I take the temporary pot out, measure it, and replace it with a fixed resistor.
 
Last edited:
I plan to add a CL80, replace the electrolytics and the rectifier, and see how it performs.

Most of that seems pretty straightforward, but I already have a question about replacing the rectifier. I was planning to use two 1.5a 1000v diodes to create a new bridge on a terminal strip for the rectifier replacement as see on this thread, but I have some GBU4G 4 amp rectifiers that I have used on a fisher and a scott.

It sounds like you're talking about two different parts of the circuit. The rectifier used in these is a 5AR4. You said you tested the 12AX7s. Did you test the 5AR4? If it tests good, keep it. The original is likely a Mullard or Amperex and it is probably fine.

The diodes would be used to replace the original selenium rectifier which is used in the bias supply and also provides DC to the heaters of the phono section 12AX7s. Take a look at the link to the thread posted above by Gadget. In it I think I posted pics and described how I did it using a pair of UF4007 diodes and a 10 ohm resistor.
 
Thanks a lot. I did not see this thread when I searched, and I'm sure it will answer most of my questions. Seems very thorough.
 
I have replaced the main 40/40/20/20 can cap with no big issues. Voltages are a little low at the four main PS can sections. All voltages here taken with variac set to 117VAC. The voltages jump to above the spec'd values when I run it on my mains voltage, which is usually around 123-124 VAC. I tried a different rectifier tube and the voltages stayed effectively the same. Only getting 409v off of the rectifier seems a little suspect at 117VAC even if these are within 5%.
Spec'd....Actual
425v.......409v
420v.......405v
350v.......341v
60v.........64v
IMG_6413.jpg

My supply voltages are a little low at the small 30/20 can cap as well, which I replaced with 33uf and 22uf 450v electrolytics under the chassis. I think this is due to the 10k 2W dropping resistor drifting up to around 11k ohms and the low incoming voltage. The final 56k dropping resistor is also a tad high at around 58.4k. Instead of 270v at the 33uf cap, I am reading 258v after the 10k dropping resistor. Instead of 250v at the 22uf cap, I am reading only 188v after the 56k dropping resistor. I have ordered replacements for these in a few different values to hopefully get my voltages in line there. I am hoping these low values are just a result of the low initial voltage from the rectifier tube, but it still seems like these are very low.
IMG_6395.jpg

After redoing the bias supply, I am getting around -20.8v which is fine with me for now. Once I add my CL-80 and get the other voltages in line, I may revisit this. I used a 10ohm resistor coming off of the new diodes to drop the voltage. I am also getting -23.5v for the heater supply for the phono tubes.
IMG_6414.jpg

As far as functionality goes, I get a low level hum in both channels. The hum is present always and does not change with volume, tone, or any input. I am thinking this is power-supply related since it isn't affected by any controls. The hum is still present if I remove the 7199 driver tubes. It is also still there if I remove the 12AX7 preamp tubes. Does this point to the output section as the culprit for the hum? I was thinking it may be a grounding issue with the 30/20 can cap replacements, so I may try to ground them to chassis instead of to the grounding bar they are wired to now. I tried to set the hum balance and a new matched quad of EL34M's, but there was no change.

There is also some extra rustling noise in the left channel only. I have swapped tubes around, cleaned sockets/pins, etc, but it stays in the left channel. Could this be related to the low voltages I am getting since the 275v feeds the non-phono preamp tube plates? Or maybe a noisy resistor in the preamp section? When I remove preamp tubes V3, V4, V5, the rustling is gone, so I am hoping getting those plate voltages up will help things.

Any comments or suggestions are welcome. I am mainly wondering how I can get my main PS voltages back up and how to go about tracing that hum.
 
What sort of current draw are you getting on the output tubes? if they're pulling too much the B+ will be low. A 10 ohm resistor between the cathode of each output tube and the ground connection allows for handy monitoring of that.

Hum with only output tubes sounds like a problem there. Could be a serious mismatch, might be grounding related if you've changed any of that. Some things are extremely picky about their ground locations. Filtering problems can also show as low B+ so confirming all your grounds are good might not be a bad idea.
 
The PS voltages are well within spec, don't worry about them at this point. Right now, lower is a good thing as long as the current draw of the output tubes is reasonable.

Voltages will be higher with modern wall voltage so there's no sense in trying to raise them now, while using 117v, because they'll just end up being that much higher later. When the unit is in actual use and it's possible you might want to lower them at that point.
 
What sort of current draw are you getting on the output tubes? if they're pulling too much the B+ will be low. A 10 ohm resistor between the cathode of each output tube and the ground connection allows for handy monitoring of that.

Hum with only output tubes sounds like a problem there. Could be a serious mismatch, might be grounding related if you've changed any of that. Some things are extremely picky about their ground locations. Filtering problems can also show as low B+ so confirming all your grounds are good might not be a bad idea.

Thanks for the suggestions. I tried two sets of output tubes, one of which was a new matched quad from a good source, so I don't think it is the tubes and it's unaffected by swapping the tubes around. I have some 10ohm resistors for measuring the current draw that I hope to install tomorrow. Excessive draw from the output tubes could explain a could of things I am seeing, so that is the next thing I will check before messing with any of my grounds. I still have one ground that I suspect that should be easy to move.
 
The PS voltages are well within spec, don't worry about them at this point. Right now, lower is a good thing as long as the current draw of the output tubes is reasonable.

Voltages will be higher with modern wall voltage so there's no sense in trying to raise them now, while using 117v, because they'll just end up being that much higher later. When the unit is in actual use and it's possible you might want to lower them at that point.

I did take the voltages running on around 124 VAC from my wall, and the first few voltages were a little high but much closer compared to the schematic until I get to the 275v and 250v supplied at the 30/20 can cap I replaced. They measure around 260v and 195v respectively on full mains voltage. The 195v is over 20% low! Those dropping resistors measured a little high, but I am not convinced it would have such a big affect on those voltages. The ground I suspect is the ground for the new 33uf and 22uf caps where the voltages are low. They're grounded to the grounding bar right now as the original multisection can was, but I may try to ground them to the chassis at the terminal strip I added and see if that does anything.
 
Ignore the slightly high dropping resistors. With the exception of the 195v section all the others are in the ballpark. Trying to exactly match the voltages in a vintage amp to the schematic is a waste of time. A tolerance of +/- 15% was normal, even when they were brand new. If my final voltages - when using modern wall voltage - are on the edge I will sometimes make adjustments.

The section that's reading only 195v supplies the phono tubes (V1 and V2) so, if you plan on using a turntable, concentrate on that section. If you have no plans to play vinyl, you can just ignore the phono stage and the section of the PS that powers it as long as the voltage of the bias supply, which also heats the phono tubes, is good. The phono section won't affect anything when you're using another input. If you're not using the phono I would suggest that you replace the Telefunken tubes in that section with some that are functional but less rare / desirable. In that case, any weak 12AX7 with functional heaters will do fine.

What schematic are you using? There are two basic designs of the early S-5000 that uses the 6BQ5 / 7189 output tubes. The only Sams Photofact I've seen is for the version that has the 12AX7s mounted at an angle with individual shields. The other version, like yours, has the 12AX7s mounted vertically and shielded by a removable panel. The Sams shows slightly different voltages so I'm guessing you're using a different schematic, unless Sams issued another one.

Make sure all the parts associated with those tubes are in spec - that's generally +/- 15% if you're using the resistance table in a Sams, which allows you to measure without removing anything. If you're using a schematic that doesn't have a similar table you'll need to check individual resistors and caps in that section. In that case the tolerances would generally be +/- 10% and you'll probably need to lift one end of the part from the circuit by desoldering it before you can get an accurate reading. It's not uncommon for old parts, especially carbon comp resistors and low voltage electrolytic caps, to drift and be out of spec. I often keep resistors that are still in spec but I always replace low voltage lytics.
 
Well, turns out I am just an idiot. I forgot how the hum balance pot worked at that there are two controls to it! I had just been turning the upper portion of it, which was allowing me to bias and balance the left channel tubes just fine, but I couldn't get my right channel tubes to balance. I wasn't turning the separate lower portion of the control. Once I figured that out, I balanced them out and the hum vanished, just like the instructions say it will do if you can read. Well here are my current draw readings after balancing each pot to -19v bias voltage at pin 2 of each output socket give or take a few tenths. I am sure I can swap tubes around to find some better balance, but here are the readings I took across the 10 ohm cathode resistors I installed:

Right channel
.266v = .027a; approx. 11.34w
.290v = .029a; approx. 12.18w

Left channel
.280v = .028a; approx. 11.76w
.312v = .031a; approx. 13.02w

Plate voltages were all within a few volts of the spec'd 420v.

The tubes still seem to be running pretty hot, correct? I am getting around -22v at the bias supply, but I am going to go ahead and change the 10 ohm dropping resistor I added after the diodes that replaced the rectifier to an 8ohm resistor I had ready for this purpose. I would rather see the current drawn around 21-22 ma, so hopefully I can get a little more bias voltage without increasing the heater voltages for the preamp tubes too much.


Ignore the slightly high dropping resistors. With the exception of the 195v section all the others are in the ballpark. Trying to exactly match the voltages in a vintage amp to the schematic is a waste of time. A tolerance of +/- 15% was normal, even when they were brand new. If my final voltages - when using modern wall voltage - are on the edge I will sometimes make adjustments.

The section that's reading only 195v supplies the phono tubes (V1 and V2) so, if you plan on using a turntable, concentrate on that section. If you have no plans to play vinyl, you can just ignore the phono stage and the section of the PS that powers it as long as the voltage of the bias supply, which also heats the phono tubes, is good. The phono section won't affect anything when you're using another input. If you're not using the phono I would suggest that you replace the Telefunken tubes in that section with some that are functional but less rare / desirable. In that case, any weak 12AX7 with functional heaters will do fine.

What schematic are you using? There are two basic designs of the early S-5000 that uses the 6BQ5 / 7189 output tubes. The only Sams Photofact I've seen is for the version that has the 12AX7s mounted at an angle with individual shields. The other version, like yours, has the 12AX7s mounted vertically and shielded by a removable panel. The Sams shows slightly different voltages so I'm guessing you're using a different schematic, unless Sams issued another one.

Make sure all the parts associated with those tubes are in spec - that's generally +/- 15% if you're using the resistance table in a Sams, which allows you to measure without removing anything. If you're using a schematic that doesn't have a similar table you'll need to check individual resistors and caps in that section. In that case the tolerances would generally be +/- 10% and you'll probably need to lift one end of the part from the circuit by desoldering it before you can get an accurate reading. It's not uncommon for old parts, especially carbon comp resistors and low voltage electrolytic caps, to drift and be out of spec. I often keep resistors that are still in spec but I always replace low voltage lytics.

I listen to vinyl almost exclusively, so I do need to get that last supply voltage in line. I will thoroughly check all of the components in that section still, but I am also going to try some different value dropping resistors in that spot before the 250v supply. I am using a factory schematic that matches this amp, not the Sams schematic that you mention, although I have cross-referenced it a couple of times. This one does not have a handy table like the Sam's schematic, so it'll be a little more work if I can't get the supply voltage in line with a new dropping resistor. Thanks for your help on this.
 
I just swapped out the 10ohm resistor for an 8ohm coming off of the rectifier diodes, and my bias voltage is now around -22.8v and the heater supply to the preamp tubes around -24v. I took some more time to balance the current draw on the output tubes. For the right channel, I can get both tubes to read around .265v across the 10ohm resistor, so 26.5 ma. For the left channel, I can get both tubes to read around .280v across the 10 ohm resistor, so 28 ma. These still seem too high to me. I don't really want to drop that 8 ohm resistor down much further. The last dropping resistor before the bias supply is reading a little high at 5.2k and is supposed to be a 4.7k. I ordered a replacement for this already that should be here today. Hopefully that helps get my bias and current back in line, but it still seems like I am drawing a lot of current. What else should I consider here?
 
Good to hear you got the hum issue sorted out.

First, are the voltages you're measuring now taken at modern wall voltage or at 117v? I'd suggest you use wall voltage if you're trying to optimize the voltages so you don't have to revisit them later. Next, I'm wondering about that last resistor and I'm a little confused. In an earlier post you said it's supposed to be 56k but measured 58.4k and that the voltage drop across it is supposed to be 20v. A later post seemed to indicate that there was supposed to be a 25v drop across it. Now you say it's a 4.7k that measures 5.2k. Perhaps you could clarify?

I figured the 56k couldn't be right. That would mean the phono tubes would only be drawing 0.357 mA. (20 / 56000 = .000357 A.) So 4.7k makes more sense. The Sams for the slant tube version shows those resistors to be a 10k 2w and a 10k 1w. It also shows a 20v drop across the last 10k so that section should be drawing 2 mA.
 
Datasheet for a 7189 says you should be about 30ma per tube so I think you're fine.

Unbalanced tubes would definitely explain the hum. Push-pull amps rely on balanced current for quiet operation. Extreme imbalance makes that not work right and you can hear how dirty the power supply actually is.
 
If you have as much voltage drop across that last resistor as you say it has to mostly be due to increased current draw in the phono stage or some other issue other than the slightly higher resistance. The difference between 4.7k and 5.2k would only account for 1v or 2v extra drop if the current draw was normal. That's based on an expected voltage drop of either 20v (4.25 mA) or 25v (5.31 mA).
 
If you have as much voltage drop across that last resistor as you say it has to mostly be due to increased current draw in the phono stage or some other issue other than the slightly higher resistance. The difference between 4.7k and 5.2k would only account for 1v or 2v extra drop if the current draw was normal. That's based on an expected voltage drop of either 20v (4.25 mA) or 25v (5.31 mA).
I see. Well my next step will be to examine the phono section and see what is going on there.
Good to hear you got the hum issue sorted out.

First, are the voltages you're measuring now taken at modern wall voltage or at 117v? I'd suggest you use wall voltage if you're trying to optimize the voltages so you don't have to revisit them later. Next, I'm wondering about that last resistor and I'm a little confused. In an earlier post you said it's supposed to be 56k but measured 58.4k and that the voltage drop across it is supposed to be 20v. A later post seemed to indicate that there was supposed to be a 25v drop across it. Now you say it's a 4.7k that measures 5.2k. Perhaps you could clarify?

I figured the 56k couldn't be right. That would mean the phono tubes would only be drawing 0.357 mA. (20 / 56000 = .000357 A.) So 4.7k makes more sense. The Sams for the slant tube version shows those resistors to be a 10k 2w and a 10k 1w. It also shows a 20v drop across the last 10k so that section should be drawing 2 mA.

I am using modern wall voltage to take these measurements now. Should have clarified. And sorry for the confusion about the resistors. There are a couple in question at a couple different parts of the schematic. Hopefully you can see from this photo.
upload_2017-6-28_17-25-30.png
One is R4, the 56K (they must have left off the K) resistor that drops the 275v supply down to 250v between the 30uf and 20uf sections of C24. I read your last post, and I think I just need to spend some time inspecting the phono section and measurements at those tube sockets.

The other is R69, the 4.7K resistor just before the -19v bias supply. It has drifted to around 5.2K. I will still replace it to see if my bias voltage rises a hair, but it sounds like my draw from my output tubes is within the acceptable limits of these tubes.
 
Datasheet for a 7189 says you should be about 30ma per tube so I think you're fine.

Unbalanced tubes would definitely explain the hum. Push-pull amps rely on balanced current for quiet operation. Extreme imbalance makes that not work right and you can hear how dirty the power supply actually is.
I'm just glad I figured it out even though it was a stupid error. I haven't worked on a ton of tube projects like this if you can't tell, but I have learned a ton about how these things work recently. I am starting to much prefer it to working on solid-state gear. I still have some rustling in the left channel that goes away with preamp tubes removed, but my next step is going to be sorting out the preamp anyways.
 
might just be a noisy tube. Try moving them one by one to the other channel to see if the noise moves. Also, clean the sockets if you have not.

Tube stuff is actually pretty simple for the most part. Much less parts than solid state amps. They can get fussy about certain stuff, but a lot of other things will work just fine with stupidly wide tolerances on voltages and components.
 
might just be a noisy tube. Try moving them one by one to the other channel to see if the noise moves. Also, clean the sockets if you have not.

Tube stuff is actually pretty simple for the most part. Much less parts than solid state amps. They can get fussy about certain stuff, but a lot of other things will work just fine with stupidly wide tolerances on voltages and components.
I've cleaned the sockets and swapped tubes around and it's still in the left channel. I need to put all of the covers back on and give it a proper test now that the output tubes are balanced. Some of the noise seemed to be affected when I was touching the knobs and chassis earlier so I'm thinking the covers could eliminate some of it.
 
Back
Top Bottom