Marshall 20W Clone Build Redplating?

[GregS]

I've just completed a scratch build of a Marshall 2061x which uses 2 EL84 power tubes in cathode bias. The amp is working and actually sounding really good. However, I have some concerns that all is not well from observing the power tubes and some of the voltage measurements I've taken in the amp.

First, I've never observed redplating before so I don't have experience in recognizing it. However, one of the power tubes has a reddish glow appearing through the plate on one side which matches the descriptions I've read. The other one seems to be on the verge of glowing red but not quite there. I measured a cathode current of 60mA on the glowing one and 51mA on the other. Do these currents sound too high for a current production Sovtek
EL84?

Also, the voltage readings I'm getting are questionable. They are all somewhat higher than the voltages listed in the plans.

B+ at the first filter cap is 387 VDC.
B+ at second filter cap (screen node) is 379 VDC.

Voltages measured at the plates and screens of the power tubes are about
375 VDC. The screens are at approximately the same voltage as the plates which doesn't seem right to me. There are no individual screen resistors in this build. There is a 470 Ohm resistor between the plate and screen nodes of the power supply. I think the resistance of the OT primary is what is lowering the plates to the same voltage as the screens.

The voltage at the cathodes of the power tubes is 13.7 VDC.

In summary what I'm asking is:

1. Does my description of what I'm seeing with the tube sound like redplating?
2. Do the cathode currents seem to high for EL84's?
3. Is it a problem that the screens are at the same voltage as the plates? If so is the solution to use individual screen resistors?

Any help would be much appreciated.

Thanks,

Greg

[olddawg]

Originally Posted by GregS

I've just completed a scratch build of a Marshall 2061x which uses 2 EL84 power tubes in cathode bias. The amp is working and actually sounding really good. However, I have some concerns that all is not well from observing the power tubes and some of the voltage measurements I've taken in the amp.

First, I've never observed redplating before so I don't have experience in recognizing it. However, one of the power tubes has a reddish glow appearing through the plate on one side which matches the descriptions I've read. The other one seems to be on the verge of glowing red but not quite there. I measured a cathode current of 60mA on the glowing one and 51mA on the other. Do these currents sound too high for a current production Sovtek
EL84?

Also, the voltage readings I'm getting are questionable. They are all somewhat higher than the voltages listed in the plans.

B+ at the first filter cap is 387 VDC.
B+ at second filter cap (screen node) is 379 VDC.

Voltages measured at the plates and screens of the power tubes are about
375 VDC. The screens are at approximately the same voltage as the plates which doesn't seem right to me. There are no individual screen resistors in this build. There is a 470 Ohm resistor between the plate and screen nodes of the power supply. I think the resistance of the OT primary is what is lowering the plates to the same voltage as the screens.

The voltage at the cathodes of the power tubes is 13.7 VDC.

In summary what I'm asking is:

1. Does my description of what I'm seeing with the tube sound like redplating?
2. Do the cathode currents seem to high for EL84's?
3. Is it a problem that the screens are at the same voltage as the plates? If so is the solution to use individual screen resistors?

Any help would be much appreciated.

Thanks,

Greg

SS or tube rectifier? What are the secondary voltage specs for your power transformer? You may want to get your B+ voltage down to around 350 vdc for EL84s. An easy way to do this is to use 10v 5 watt zeners in series.

[tubeswell]

Red plating is a sign that too much current is flowing through the plate, so you could decrease the plate voltage by lowering the B+ as olddawg says,

...or you could increase the bias voltage, either by upping the value of the cathode resistor (if its cathode biased), or making the grid voltage more negative (if its fixed bias).

[Chuck H]

I think more needs to be done than lowering the plate voltage to 350V. Right now those poor el84's are dissapating 18 and 21 watts. For a total dissapation at idle of 39 watts for a 20 watt amp. Thats WAAAAYY too much. Even when most Marshall 18 and 20 watt clones are running properly they stress the power tubes.

If you try to reduce current to an acceptible level by raising the value of the cathode resistor your plate voltage will increase a lot. So...

You should either replace the power transformer or add a zener voltage reduction circuit as suggested.

Increase the size of the cathode resistor.

Increase the screen node dropping resistor to 1k or 2.2k

Add individual screen resistors of 470 ohms.

I would say that in general you want a plate voltage under 360V. You want the current through each tube to be somewhere between 30mA and 40mA. The actual current will depend on the actual voltage and the desired class of operation. The class of operation will be determined by the current and,to some extent, the amount of drive to the power tube grids.

You could just reduce voltage and current. But the extra reduction on the screens and the added resistors will improve tube life a bunch. There would be a change in tone. But it probably won't be an objectionable change.

Chuck

[GregS]

Thanks for the replies.

SS or tube rectifier? What are the secondary voltage specs for your power transformer?

It's a solid state rectifier with two IN4007 diodes. The secondary voltage specs for the power transformer are 320V - 0V - 320V. I measured the voltage across them to be 655 VAC out of circuit so it seems that it's operating within spec.

Since I'm following a current production amp design (the re-issue Marshall handwired series 2061x model) and am using all the specified part values in the plans, it doesn't make sense to me that the design could be so flawed - otherwise the production Marshall amp would have the same problems. For this reason I'm inclined to look for simpler problems to start with rather than playing around with the design.

I did not deviate from the plans other than to have the cathodes pass through 1 Ohm resistors across the test points before joining up and then going to the 120 Ohm cathode resistor. The only other difference is that I combined multiple resistors in series (in the power supply) to get the exact specified values where I didn't have a single resistor of that value. I'm kind of new to this as this is only my second build but I don't think these minor deviations should change anything significantly???

There is some possibility that there is a mistake in one or more of the part values in the plans but I'm doubtful of this. The voltages I'm getting at the power supply nodes are around 30V higher than those on the schematic. However, they are about the same as those in the schematics of similar old Marshall 20W amps that I found on Schematic Heaven.

There is a 22 Ohm resistor between the rectifier and the first supply node. If I used a higher value resistor here that should drop all the supply voltages into the target range???

Also, as I mentioned in my first post, one of the tubes has a spot of orange glow about 1/2" in diameter on one side. Does that description sound like redplating?

Thanks for the help,

Greg

[tubeswell]

Originally Posted by GregS

The voltages I'm getting at the power supply nodes are around 30V higher than those on the schematic. However, they are about the same as those in the schematics of similar old Marshall 20W amps that I found on Schematic Heaven.

Depends on your wall voltage. Older amps were made to run in days of slightly lower wall voltages. You need to re-bias your tubes

Originally Posted by GregS

There is a 22 Ohm resistor between the rectifier and the first supply node. If I used a higher value resistor here that should drop all the supply voltages into the target range???

A sag resistor? Yes if you increase the R value, you will lower the B+ but that is a hard way to do it (the resistor also needs to have a good power rating to withstand the heat it is having to dissipate. You could also use big chunky limiter resistors between the PT HT secondaries and the rectifier diodes. See the valve wizard for a detailed explanation of how to calculate the right size.

Originally Posted by GregS

Also, as I mentioned in my first post, one of the tubes has a spot of orange glow about 1/2" in diameter on one side. Does that description sound like redplating?

Yes. A 1/2" diam orange or red glow on the plate is 'redplating'. Eventually the tube will self destruct unless you change the operating conditions.

Also what tubes are you using? (I have found out that some modern batches of Sovtek EL84s are not as robust as earlier made ones - especially the newer ones that come in the mainly red boxes, and I have had redplating problems in the recent past with the newer tubes in situations where older Sovtek EL84s have worked without issues)

[GregS]

Thanks for the reply Tubeswell.

A sag resistor? Yes if you increase the R value, you will lower the B+ but that is a hard way to do it (the resistor also needs to have a good power rating to withstand the heat it is having to dissipate. You could also use big chunky limiter resistors between the PT HT secondaries and the rectifier diodes. See the valve wizard for a detailed explanation of how to calculate the right size.

Great! That's what I wanted to hear! I am using big chunky 10W wire wound resistors for that resistor and the one between the OT center tap and screen nodes already so that shouldn't be a problem.

Depends on your wall voltage. Older amps were made to run in days of slightly lower wall voltages. You need to re-bias your tubes

Actually, I think I know why the voltages are high now despite following the plans. The power transformer I'm using (Mercury Magnetics P20JM Tone Clone) is not correctly spec'd to the circuit. The schematic was taken from the modern re-issue version of the amp and I suspect the voltage measurements on the schematic were also taken from this amp. The Marshall re-issue does not use the same PT I'm using which is likely the reason for the voltage descrepancy. The lower voltages of the re-issue are probably necessary to accomodate modern less robust tubes. The PT I'm using is supposed to be an exact clone of the one used in the Marshall amp from 1974 which explains why my voltages more closely match the ones of the old amp. The plans listed the PT I'm using as one of the possible options. There are some other differences between the old amp and the re-issue relevant to the biasing of the power tubes as well. The old amp uses a slightly greater value cathode resistor of 125 Ohms instead of 120 Ohms. The old amp has a larger value resistor between the plate and screen nodes of 750 Ohms compared with 470 Ohms.

It would be a real pain to replace the PT at this point as I'd essentially have to rebuild the amp, not to mention the expense of this transformer. I have to try to make it work with this one. Hopefully, increasing the value of that sag resistor will solve the problem.

Initially when I tried it out I had it plugged into my lightbulb cord and all the voltages were low as a result. It might be worth a test to see if the redplating problem goes away when I use it as it would confirm the high supply voltages are the root problem.

Also what tubes are you using? (I have found out that some modern batches of Sovtek EL84s are not as robust as earlier made ones - especially the newer ones that come in the mainly red boxes, and I have had redplating problems in the recent past with the newer tubes in situations where older Sovtek EL84s have worked without issues)

They're new Sovtek EL84s. However, I think the operating conditions are not right for any tube at the moment.

One more thing: when I played the amp it stayed clean all the way up to high volume levels and even at full volume there was not much distortion. I believe this is a symptom of the biasing problem I have in the power section.

Thanks for the help from everyone. I'll try the fix I mentioned and report back.

Greg

[Chuck H]

You should re-read my above post. You don't seem like you would have any trouble wrapping your head around the concepts therein. Your amp cannot be made to work properly as it is and requires change to do so. My latest endeavor is a 2x el84 amp. The PT I'm using has a 275-0-275 secondary and a 115V primary. This essentially renders it a 290-0-290 secondary with modern wall voltage. I'm using a 130 ohm cathode resistor (about 11 volts on it) and getting 355 volts on the plates. Thats a total dissapation of 15 watts per tube. If you subtract for screen current and the voltage on the cathode resistor I'm still at about 100% dissapation. As a result I'm using a 2.2k resistor between the plate and screen nodes and it reduces the screens to 330 volts. I am also using new production Sovtek el84's. But mine are matched better than yours. Compared to what you have going on this is conservative, and this amp will still eat up a pair of power tubes in less than a year with moderate use. If you don't make some significant changes your amp is not only NOT what was intended by the MFG, but it's just a tube eating monster that you will have to feed and feed.

I would suggest one or two reverse biased zeners on the center tap of the PT B+ secondary, a larger cathode resistor and at least 1k between the plate and screen B+ nodes. Individual screen resistors are just a good idea as most amps that didn't have them were eventually released with them. There's got to be a reason for that. This is an easy and affordable solution.

Chuck

[GregS]

You should re-read my above post. You don't seem like you would have any trouble wrapping your head around the concepts therein. Your amp cannot be made to work properly as it is and requires change to do so.

Thanks for your posts Chuck. They are informative and I have taken your suggestions into consideration. I realize the amp cannot work as is. I'm only turning it on long enough to take readings to troubleshoot and shutting it off before the tubes melt.

If you don't make some significant changes your amp is not only NOT what was intended by the MFG, but it's just a tube eating monster that you will have to feed and feed.

The first significant change I need to make is to bring all the supply voltages down to those specified in the schematic. My supply voltages are in correct proportion to each other as per the schematic. They are all about 30V too high. My plan to get the voltages down is to use a larger value resistor between the rectifier and the first supply node. Once this is accomplished, hopefully the amp will be operating according to the design. If it is still redplating I'll have no choice but to make some of the other changes. If it is working as designed but eating power tubes it will depend on what compromise the person I'm building it for wants to make - either have it exactly like the Marshall but replace tubes constantly, or have it easier on tubes but not exactly like the Marshall.

I would suggest one or two reverse biased zeners on the center tap of the PT B+ secondary, a larger cathode resistor and at least 1k between the plate and screen B+ nodes. Individual screen resistors are just a good idea as most amps that didn't have them were eventually released with them. There's got to be a reason for that. This is an easy and affordable solution.

I'm not familiar with the zener diode thing off hand but may be able to find it in my reference material. Is the purpose of this change to lower the B+ voltages?

The old version of the amp had a 750 Ohm resistor between the plate and screen nodes which is closer to the 1K that you suggest. The 470 Ohm resistor of the new re-issue version does not make sense to me as the voltage at the plate and screen wind up being about the same which I don't think is good. Probably the OT used in the re-issue isn't spec'd the same as the MM 20W Tone Clone I'm using. The primary may have lower resistance resulting in the plate having higher voltage than the screen in the re-issue. Since the transformers I'm using are clones of the originals from the old version of the amp, using a 750 Ohm resistor here should put the plate and screen voltages in the same proportion as they are in the old amp. Using a 1K as you suggest may be a better idea though.

Thanks again for your help.

Greg

[Chuck H]

Originally Posted by GregS

I'm not familiar with the zener diode thing off hand but may be able to find it in my reference material. Is the purpose of this change to lower the B+ voltages?

Yes. Placing a reverse biased zener between the PT HV secondary CT and ground will lower the B+ by the Zener voltage. Sometimes it's best to use two in series to improve the wattage handling and reduce the heat dissapated by just one. They are big chassis mount diodes that get hot so it's best to mount them away from any filter caps or other heat sensitive components.

To reduce the B+ 30V with a resistor you will need to add at about 150 ohms. So you could try a 180 ohm resistor. But... Because it will be dropping more volts it will be dissapating more power as heat. You will need a resistor that can handle the watts. Let's assume your dropping 35 volts total. Now consider that with the tubes drawing current that amount will go up. I'd ballpark it at 55 or 60 volts dropped across that resistor under load. So...

(Volts divided by resistance will give the current. Current multiplied by voltage will give you watts) 60/180*60 = 19.99

A 25 watter would not allow an appropriate safety margin so you'll need a 50 watt resistor. I'm using a 150 ohm 50 watt aluminum housed chassis mount resistor (the gold ones with fins you see around) in my 2Xel84 amp with a plate voltage of 355. My resistor gets pretty hot when I run the amp wide open for awhile (which is often ). So I'll recommend the 50 watt resistor. Do otherwise at your own peril.

Also, consider that by adding all that extra resistance to the rail you will be increasing the sag envelope a lot. It's no different than if you used a rectifier tube instead of diodes. So any tightness from the diode rectifier will be lost and the feel of the amp will be different from an original in this regard. Probably more different than if you implemented the changes I suggested. But IME (And I do have a similar amp to what you propose) the amp won't get mushy or squishy. In fact I think mine has a very good envelope and balance of attack and compression. So it's your call. Either way works fine.

Chuck

[tedmich]

my PV Bravo loves ~400VDC B+, fixed bias ~18v with the JJ EL84s, but the zener (or zener/mosfet) trick should put the B+ down where this design likes it.

[GregS]

(Volts divided by resistance will give the current. Current multiplied by voltage will give you watts) 60/180*60 = 19.99

A 25 watter would not allow an appropriate safety margin so you'll need a 50 watt resistor. I'm using a 150 ohm 50 watt aluminum housed chassis mount resistor (the gold ones with fins you see around) in my 2Xel84 amp with a plate voltage of 355. My resistor gets pretty hot when I run the amp wide open for awhile (which is often ). So I'll recommend the 50 watt resistor. Do otherwise at your own peril.

I was just at a local electronics store today and bought a number of 10W resistors to try out to drop the voltage the target amount. I wasn't thinking that a voltage drop that large would require a resistor that could withstand considerably more power but makes sense to me now. Good thing I checked out this thread again before doing it! I did notice they had the the higher wattage gold ones with fins that you mention. However, I don't like the idea of having the effect of a tube rectifier. If I understand correctly, the zener diode method will drop my B+ as needed without losing the tightness of the diode rectifier. I'll look into this option.

Thanks again for your help.

Greg

[Chuck H]

As a temporary experiment you could series two of your ten watters that add up to about 180 ohms. The resistors won't have even current across them if they are mis-matched but I don't think they would blow right off, and not necessarily at all. It would be a good way to find out how it sounds this way.

There are a lot of threads about the zener voltage reduction here. If you do a search you'll find out all about it.

Chuck

[tubeswell]

Or you could use 4 x 560R 10W in parallel for about 140R @ 40W

With the zenering method you need an HT winding that has a centre tap - I wasn't sue if your PT's HT winding had a centre tap or not. The zener effectively lowers the 'starting' AC voltage on the HT relative to the ground potential by shifting the HT winding's voltage reference point below ground potential. So a 30V reverse-biased zener on the HT would bring a 330-0-330VAC winding down to 300-0-300VAC relative to ground potential, by setting the "0" point at -30V. The diode has to be strong enough to withstand the total maximum current draw experienced on the HT winding - so multiply the total maximum current draw x 2, and then multiply that by the voltage dropped by the zener to get the required power rating for the zener.

If you can't find a 20W 30V zener, don't worry string 3 x 12V 5W or 4 x 9V 5W reverse biased zeners in series on a tag-board and you should be alright (20W is overkill anyway in most situations calling for this)

[GregS]

I think for the moment I'll try the resistor method and see what happens since I already have them . I don't think I can get ahold of the necessary zener diodes that quickly.

I've tried to calculate what value resistor value I need and how much power it needs to be able to handle based on voltage mearsurements I've taken in the amp with the tubes in drawing current. However, I want to be sure I haven't missed something before assuming this is correct.

There is an existing 22 Ohm resistor between the rectifier and the 1st supply node. I measured 390V on the rectifier side of the resistor and 387V on the other side for a voltage drop of around 3V. This means that the current flowing through the resistor is 3V/22R = 0.136A.

Assuming 0.136A current regardless of what resistor value is used here (I'm doubting this is correct) and I want to drop the voltage 45V that would mean 45V/0.136A = 330R. If I use a 330R resistor it will have to be able to handle 45V x 0.136A = 6.12W. Does this make sense or am I way off here? I'd rather not find out the hard way that my logic is wrong.

Incidentally I also did some tests with the amp going through my lightbulb cord. This drops the voltage at my first supply node down to 287V from 387V. Certainly no redplating occurs under these conditions. However, the tube that was redplating has a cathode current of 47.9 mA and the other one 22.9 mA under these conditions. It seems they've drifted quite far apart. I suppose the one that was redplating has already sustained significant deterioration just from the brief time I've had it on during my tests. To be clear, I'm not suggesting the lightbulb cord as a solution to my problem.

Greg

[Chuck H]

Originally Posted by GregS

Assuming 0.136A current regardless of what resistor value is used here (I'm doubting this is correct) and I want to drop the voltage 45V that would mean 45V/0.136A = 330R. If I use a 330R resistor it will have to be able to handle 45V x 0.136A = 6.12W. Does this make sense or am I way off here? I'd rather not find out the hard way that my logic is wrong.

Your logic is fine. But incomplete. The larger resistor will have higher current across it. I can't say exactly why, it's probably something to do with the plate and OT primary resistance being in a parallel circuit to the B+ rail. So maybe the closer you get to that resistance (the plate and OT primary resistance) with the dropping resistor, the nearer the two current levels match. At any rate, though I can't say exactly why, I can say that a larger resistor will have more current across it. But by the same token that means you don't need as high value a resistor to drop the volts. Even though I can't pin down the exact cause, I do own a working example of this. In my cathode biased 2xel84 amp I have 398 volts at the rectifier. I'm using a 150 ohm dropping resistor and it drops about 43 volts from the supply with the tubes set to idle at near 100% dissapation. So...

If my 150 ohm resistor (operating under very similar circumstances to yours) is dropping 43 volts. 43/150*43=12.33 watts

That's at idle. In use the current across that resistor will be on the order of 150% of 12.33 watts. 12.33*150%=18.495 watts.

With safety margin that's a 25 watt resistor. But wait... There's more...

Your amp has a higher voltage PT secondary than mine. So how come we both have 390 +/- volts at the rectifier??? It's because your amp is drawing excessive current right now and this is loading the HV rail. When you re-bias your power tubes (with a bigger cathode resistor) to draw 100% (or preferably a little less) of their rated max dissapation, your voltage is going to go up and you'll be back to square one. So you'll need a bigger dropping resistor than I am using, but probably not 330 ohms. I guessed at 180 ohms and I think that's a good start. I bet it'll be 180 to 220 ohms. So...

The higher value resistor again will require a slightly higher rating. Now the 25 watt won't offer much safety margin and live a hot, hard life close to it's max rating much of the time. Since most resistor ratings jump from 25 watts straight to 50 watts, I say go with 50 watts. You could use two 15 watt or three 10 watt resistors in series or parallel too. But since one resistor looks tidier and is easier to wire in, I say just go with the 50 watt. And as I said before, do otherwise at your own peril.

Chuck

[GregS]

Your logic is fine. But incomplete. The larger resistor will have higher current across it.

I suspected that something wasn't right in my assumptions. I'll try out around 180 Ohms as you suggest for a start and see where that gets. Also, it's good to know what is and isn't a safe power rating to use.

Thanks again,

Greg

[olddawg]

Another choice you have is to use an EZ81 tube rectifier. That will bring the B+ down considerably. If it has to be SS rectification, as I stated before, stringing 10v 5w zeners works fine to lower the B+ (It also works for tube rectified B+). The best thing to do is mount the zeners on a separate piece of perf board in a zig zag arrangement mounted away from the chassis where they can breath. They get tend to get rather hot.

[Chuck H]

The good thing about stringing 5watt diodes is that they come in the standard DO-XXX case type so there's no chassis mounting. Mouser has 5watt 14V DO-201 zeners for 52 cents. But try to get some other stuff you need while your at it (gun tips, scope probes, shrink tube, lead wire, etc.) because their minimum shipping charge is about $10.00

Chuck

[Diablo]

Here's a link with some good information on Zener diodes....

https://taweber.powweb.com/store/vdump.htm

[GregS]

Great. Thanks for the info from all.

Greg

[GregS]

I ordered the VDump kit from the Weber site. I chose a 40V Zener rated for 25W (only this power rating was available).

After installation all was great at first. The voltages were dropped exactly to spec, there was no more redplating, the amp sounded great. When trying it out again today an unfortunate incident happened that damaged the zener I believe. While I was was inserting the DMM probes into the jacks I have to test the cathode current of the power tubes (effectively measuring the voltage drop across 1 Ohm resistors before the main cathode bias resistor) there was a clicking in the speaker, it seemed the lights dimmed in the room slightly, and power was gone to the amp. It turns out that the AC cord wasn't plugged into its recepatacle properly and was probably dislodged when I was trying to measure the cathode current. Once put back in properly the amp came on again, however, now one of the tubes was redplating again. When re-checking the voltages they were all high again although not quite as high as before installing the zener.

The information on the Weber site concerning the zener mentions that if your amp has a solid state rectifier and a very high value of filter capacitance you can get a large current spike when the amp is turned on. He mentions putting a NTC in-rush thermistor in series with the diode to protect it. I didn't get this thermistor as I wasn't sure what it was, couldn't find it on the site, and thought it probably wasn't necessary for a 20W amp. The amp does have a solid state rectifier and uses 50uF filter caps between ps nodes.

I suspect what happened was the power was turned on and off rapidly by the power cord not being in properly which caused a large current spike that has damaged my zener so that it is no longer dropping as much voltage as it should. I did have a 500 mA slo blo fuse in the HV CT that did not blow so I'm not totally certain of this.

Anyway, the point of this post is does my theory of what has happened make sense or could the problem be that something else is damaged??? If what I think happened makes total sense then I'll have to go ahead and order another zener and this time get this thermistor thing as well to prevent this happening again.

Thanks for any input,

Greg

[Diablo]

Yes, from your description that sounds like you had a voltage spike that killed the zener. Weber has the NTC inrush current limiter thermistor listed on the bottom of the page with diodes.

https://taweber.powweb.com/store/diodeord.htm

[GregS]

Thanks for the information on where to find that current limiting device Diablo. I'll be sure to get it along with the replacement zener diode.

I'm glad this happened during my test while I was watching things closely. Otherwise a lot more damage could have been done.

Greg