Replacing filter capacitors in JCM 800 with vertical inputs

[martin manning]

The amp is working now, so it would make sense to measure voltages and bias current in its current state, before doing the cap replacements, so you have a baseline to compare against. I would expect some changes after the cap job, like the B+ coming up a bit, and the need to adjust the bias voltage.

[shattered]

Put your clip-lead on the 50uf section of the last capacitor that leads to the preamp plate resistors. If you put it on pin 1 of V1, you'll be there all day waiting for it to bleed down.

One thing to be very safety conscious of is the main filters may retain a significant charge if the HT fuse is blown or removed, while the rest may be reading 0v.

I also have a tool for safely discharging HV caps, but the one under the PCB seems a bit tricky. I plan to remove the nut from the chasis and hopefully the cables will be long enough to allow me to access to the terminals . Yeah, I've seen them going up in voltage after discharged.

Fuses are correct, thanks for the heads up.

2. I don't know what the hardware size is, but I'd suspect metric. Maybe there's a Harbor Freight or similar tool store nearby? I always like a good excuse to buy another tool.

3. I think the other folks covered this well already, as those are the bias supply filters. I wouldn't bother replacing them with film caps. A decent quality electrolytic would do quite well. You'll pay a big premium for axial leaded parts. It is worth looking to see if a radial electrolytic will fit there with a little fiddling. Amplified Parts has an axial 8uf @ 150v that will work fine there, 105c rated and only $1.05 each.

4. The 2203 preamp (in the above listed schematic) section only has one cathode bypass at V1A, a 0.68uf. It is probably a film cap already, so no need to replace it. There are electrolytics in that value, so it is possible. If it is electrolytic, i'd replace it with a 0.68uf film.

Yeah, I got confused, there are only 6 big 50+50µF 500V and two 10µF 160V.

I plan to buy Sprague Atom 10µF 150V at Banzai music, where I'll also buy the F&T 50+50µF 500V and some micalex noval sockets as spares.

The amp is working now, so it would make sense to measure voltages and bias current in its current state, before doing the cap replacements, so you have a baseline to compare against. I would expect some changes after the cap job, like the B+ coming up a bit, and the need to adjust the bias voltage.

It makes sense.

Thank you all!

[shattered]

Before you put the amp back on leave do not put the output tubes back in, turn the power switch on with the standby switch off and then confirm that on pin 5 of each output tube that you read at least a negative 53 volts for starters if your output tubs are 6550 types .

Then your ok to pop the output tubes back in

Did I get this right?

"Power switch on" light on, the amp is powered. Heaters are on but no B+ voltage yet since Standby button is in "0" position.
"Standby switch off" means standby switch is pressed down. This is the position where the amp produces sound.

As suggested in other posts I plan to take some measures before doing any change in the circuit. So, with both switches up, first I press the Power button, wait for the tubes to heat, then press Standby so B+ voltage is applied. I plan to measure voltages at pin 5 of each output tube, B+ and bias current before the recap. Could you confirm it is safe for the amp to measure voltage at pin 5 with the tubes inserted? volume set to minimum, of course with a speaker connected.

Once I do the recap I follow your procedure, measuring voltages before installing the tubes again.

However... I'm still not sure if I should follow this procedure, more complex than I imagined, versus just using the amp meanwhile it sounds fine — I use it gently at very low volumes — and change the electrolytics when I observe any problem like hum or pops. Even following safe procedures like wearing rubber soles and keeping one hand in the pocket and the probe tips being insulated up to the tip I'm afraid of making a short.

[nuke]

Have you done this work before?

[shattered]

Have you done this work before?

No, just bias adjustment using the probe and checking plate voltage with a voltmeter attaching an alligator clip to pin 3, never poked inside a live amp apart from turning the bias pot with a non conductive screwdriver. I try to work as foolproof as possible.

I recently made a successful repair of my Laney that intermitently dropped voume and sounded distorted, I replaced 4 capacitors that looked very suspect with signs of electrolyte leak and an octal socket that was partially carbonized (a previous owner retubed but didn't bother to replace the socket too), retubed and also upgraded screen resistors from 470 ohm to 1K. The old tubes were 5881 with around 10mA difference in bias current and one of them looked suspect. Replacing them with fresh EL34 got rid of scratchy sound and jumpy volume in the lower master volume settings (I initially thought it was the potentiometer).

I have experience soldering both THT and SMD and did a good number of recaps to other equipment (for me, never for others) with success. I know the basics for my own safety but I'm afraid to create a problem fixing a problem that doesn't exist... specially in this amp.

[martin manning]

Sounds like you have the experience to take this on. Poking around tube sockets can be risky, but you can often stay out of trouble by probing on the board rather than at the socket lugs. For example checking for bias voltage can be done by probing the bias supply caps or the 220k (or 150k in your case) power tube grid leak resistors. It still might be a good idea to verify that the bias voltage is actually getting to the sockets if there is any doubt. Of course, clip the ground lead to the chassis so you can probe with one hand. A set of probes with retractable sheaths that leave only the tip of the metal part exposed are useful too.

Two of the most confusing terms in this business are standby on vs. standby off, and higher bias voltage vs. lower.
I am careful to say "standby on standby" when I mean the amp is silent and (usually) there is no HV on the tube plates.
For bias voltage, I use the term "more negative" when the voltage is lower, for example -50V vs. -45V.

[Stevem]

You may be your own worst enemy.

5881s and EL34s require two very different bias voltage settings .

Did you make the needed change?

[shattered]

Thanks, Martin. I'm still preparing mentally for this work and learning as much as I can. I want to enjoy the amp without the worries.

Would be sensible to put the probe with the current tubes, then decrease bias current around 45-50% dissipation before doing the recap? Just in case the bias current goes upwards after the operation and taking into account that I'll use the same tubes. I'm also afraid if set the most negative voltage (coldest setting), wouldn't B+ increase, risking the circuit? And removing the tubes will also increase it.

You may be your own worst enemy.

5881s and EL34s require two very different bias voltage settings .

Did you make the needed change?

No problem with that, the Laney amp (GH50L) has a switch to select between 6L6 and EL34 so it changes the bias range from colder to hotter. In the schematic they specify a 1K5 screen grid resistor, this one is mid 90s and came equipped with 470Ω and 5881 tubes, so the only mod I did was upgrade it to 1K as a balance between keeping it as stock as possible and tube safety.

Before adjusting bias I measured wall voltage (234V), now it's sitting at 37.5mA bias current according to the Bias Master and 438V plate voltage, so it should be less than 65% taking into account this measure includes the screen current. I find it sounds its best at ~63% dissipation, with a colder setting it sounds less alive and dynamic.

[martin manning]

Would be sensible to put the probe with the current tubes, then decrease bias current around 20% less than the current setting before doing the recap? Just in case the bias current goes upwards after the operation and taking into account that I'll use the same tubes. I want to do it as simple and risk free as possible.

No need to over think it, just turn the bias voltage as low as it will go (most negative) before you put the tubes in and apply power. If you can't monitor both the plate voltage and the plate current as you bring the idle current up, measure the B+ with the idle current as low as it will go and calculate a target current. Dial that in, recheck the plate voltage (it will come down), and repeat.

I would treat the start-up after you change the caps like a new build. You could read up on that, but basically it should go like this:
Start with no tubes installed and power up in standby, preferably using a current limiter.
Check for bias voltage, dial that down as low as it will go, then switch to play and check for B+ on all of the filters.
If that's all good, switch the power off and drain the filters using a clip lead from a V1 plate lead to ground. Yes it will take a bit of time; monitor the voltage as it is going down.
Put the tubes in, connect a load to the output, and power up again, first on standby, then switch to play.
Still good?
Bypass the limiter, measure B+ and calculate target idle current.
Bring the bias voltage up to hit the target current and check to see if the output tubes are reasonably well balanced.
Measure B+ again, recalculate a new target current, and adjust bias again.
Remeasure B+, but usually one iteration is enough.

[shattered]

Thank you so much. Will look into the light bulb current limiter.

[shattered]

3. I think the other folks covered this well already, as those are the bias supply filters. I wouldn't bother replacing them with film caps. A decent quality electrolytic would do quite well. You'll pay a big premium for axial leaded parts. It is worth looking to see if a radial electrolytic will fit there with a little fiddling. Amplified Parts has an axial 8uf @ 150v that will work fine there, 105c rated and only $1.05 each.

Oh, missed this part! Thanks for the suggestion.

I'll probably go that route using Panasonic NHG 10µF 160V from RS. They are General Purpose and rated 105C, AEQ-Q200. I have more trust in these ones than the Sprague Atoms.

Being careful bending the pins with a plier and maybe applying some electronics sealer like Chipquick EGS10C-20G to attach them to the PCB at the base I hope it will be fine. Dielectric strength is 460 V/mil (18 kV/mm), would be a problem if this product touches both legs?

[nuke]

Don't silicone the leads, just the body to the board. It only takes a little dab to secure them. Use an electronics grade silicone, which means it doesn't contain acids or corrosives. (most silicone sealants release acetic acid while curing). ASI 388 works well, you can get it easily on Amazon.

Seems like this might be an experience adventure for you, which is good. Consider doing one stage at a time. Do the main filter caps first, test it. Then the next filter capacitor down the line, test that. Then so forth till you get them all.

Since I've done so many of them over the years, I just do them all at once. But, there's no shame in taking one's time. When I build stuff up from scratch-new, I usually power things on a stage at time to each stage function before moving to the next.

On fitting radials vs. axials, you'll have to decide whether it is worth the effort and whether it will be secure once installed.

If the leads on the radial cap are long enough, you may be able to fold the negative lead over the side to the top side of the can and slip heat shrink tubing over it to hold it in place. That's a common method, if the leads are long enough.

[shattered]

While I wait for the components to arrive I've been experimenting with my less valuable amps, powering them (no tubes and no speakers), checking voltages and using the lightbulb current limiter to have an idea on what to expect. All worked as expected and now I'm more confident. Thank you!

In order to check the socket pins safely I attached the negative tip with an alligator clip to the sleeve of a cable in the input jack, and probed with the red tip on pins 3 and 5 of the (removed) power tubes from the top. This way there are no chance of accidentally shorting pins.

Regarding the speaker output jacks:

No need for a speaker to be plugged in also.

These amps have shorting type output jacks so even if the amp is powered up and played the self shorted jack will prevent any harm to the output transformer.

They are wired in parallel so if there is only one speaker connected the output would be shorted by the other connector. Am I missing something?

[timrobertson100]

They are wired in parallel so if there is only one speaker connected the output would be shorted by the other connector. Am I missing something?

Only one wired to short?

(The JCMs I’ve had haven’t been wired to short to ground)

[shattered]

They are wired in parallel so if there is only one speaker connected the output would be shorted by the other connector. Am I missing something?

Only one wired to short?

(The JCMs I’ve had haven’t been wired to short to ground)

I'm talking about the speaker jack outputs. I think they should be non-shorting since they are wired in parallel... otherwise at least one of them would short if it has less than two speakers connected.