Sluckey's Dual Marshall Build series

[pompeiisneaks]

Here's the first video, comments, suggestions, etc, please fire away!

~Phil

[pompeiisneaks]

Video 2:

~Phil

[pompeiisneaks]

Video 3:

Enjoy! :D

~Phil

[sluckey]

You mentioned that your PT requires a FWB recto. Easy to do with that layout as far as the B+ is concerned. ***BUT***, don't forget. You gotta change the bias supply also! That will require a bit more thought.

[pompeiisneaks]

You mentioned that your PT requires a FWB recto. Easy to do with that layout as far as the B+ is concerned. ***BUT***, don't forget. You gotta change the bias supply also! That will require a bit more thought.

Yeah I was pretty sure of that too, but doesn't that just require a slightly larger dropper resistor?  it should basically come in at a much higher C- voltage and I'll need to drop it down a bit more?  Either the 15k increased to drop the voltage a bit,  or the 47k decreased to allow more to shunt to ground? no?

~Phil
 

[sluckey]

You must use a capacitor between the PT and the bias diode. Look at my Amp Scrapbook to understand why. This is the easiest way to modify that layout for the correct circuit...

[pompeiisneaks]

Oh ok sounds good, I'm glad I bought multiples of those when I first did the AC100 in this thing. :D

~Phil

[sluckey]

I can remove replies 3, 4, 5, 6, and 7 if you want to keep this thread clean. I just wanted you to be sure what you had to do to the bias circuit. LMK.

[pompeiisneaks]

nope I appreciate all the comments, they helped me and may help some long lost future builder :D

~Phil

[pompeiisneaks]

So now I have to double check something. 

You put a 56k instead of a 220k in the first resistor to earth in that chain, on purpose I gather, does it also need to be 2W?  If my math is right, if its 56k and it dropped about the whole max 100V that would be seen there, it would be only .18 watt (100/56000) * 100  Seems a half watter would work.  I think it's more around 70ish VDC about there, though...  (since it's not an inline dropper before rectification?).   or am I up in the night?

~Phil

[sluckey]

The voltage across that 56K resistor is somewhere between 200V_pp and 400V_pp depending upon the output from your PT. It's a crazy looking 60Hz AC waveform (not a clean sine wave). Ain't no DC across that resistor. If you stretch your imagination and say the waveform is a sine wave and also say it's 400V_pp, then it would be 200V_peak or 141V_rms. Power would be 141*141/56000 = .36W. Double that for safety and buy the next higher standard wattage resistor which would be 1 watt.

Marshall JCM900s used a 56K 1/4W if you believe note 7 that says all resistors 1/4 watt unless stated otherwise. I really like the insurance of a metal oxide 56K 2W or 3W. Bias supply is cheap and simple but a lot of expensive stuff is counting on it being bullet proof.

[pompeiisneaks]

Oh that's true, the rectifier only chops half of the waveform, its' the capacitors that tend to straighten the line and lower the voltage, but aren't those caps 100V rated?   I guess you're right, though, making the bias supply rock solid is smart.  I'll go to the local shop and buy a 2 or 3 watter.  I just had ordered the 220k one that was on the original sheet not yet realizing the needed change here. 

Sadly, the X1 caps I have are a bit wider than those eyelet holes, so I'm trying to figure out the best way to put it in.  It also has super short leads like the style for a PCB, so that too makes them a bit of a PITA :D  I'll get it sorted though.

~Phil

[sluckey]

Oh that's true, the rectifier only chops half of the waveform, its' the capacitors that tend to straighten the line and lower the voltage, but aren't those caps 100V rated?

Best not to think that "the rectifier only chops half of the waveform". That's only true if the AC waveform swings equally above and below zero volts. That's not the case with this circuit. In this circuit, the voltage swings big positive above zero volts, but it swings only a little bit negative below zero volts. The diode blocks (chops) every bit of the positive and only passes the negative on to the caps. Then the caps charge to the peak of the negative pulses. So, even though you have an AC waveform that swings several hundred volts between the max positive and the max negative peaks, if the max negative peak is only 75V, then that's all that the diode will allow to pass on to the cap. Hence the 100V rating.

I think the more words I use, the more confusing this may seem, so look at page 6 of my amp scrapbook. Hopefully the scope pics should make this crystal clear.

You need to stay with the .047µF value for the class x cap. It's not a coincidence that the capacitive reactance of that value cap is 56K, same as the resistor. I would suggest using 10µF @ 150V filter caps though.

[pompeiisneaks]

Oh that's true, the rectifier only chops half of the waveform, its' the capacitors that tend to straighten the line and lower the voltage, but aren't those caps 100V rated?

Best not to think that "the rectifier only chops half of the waveform". That's only true if the AC waveform swings equally above and below zero volts. That's not the case with this circuit. In this circuit, the voltage swings big positive above zero volts, but it swings only a little bit negative below zero volts. The diode blocks (chops) every bit of the positive and only passes the negative on to the caps. Then the caps charge to the peak of the negative pulses. So, even though you have an AC waveform that swings several hundred volts between the max positive and the max negative peaks, if the max negative peak is only 75V, then that's all that the diode will allow to pass on to the cap. Hence the 100V rating.

I think the more words I use, the more confusing this may seem, so look at page 6 of my amp scrapbook. Hopefully the scope pics should make this crystal clear.

You need to stay with the .047µF value for the class x cap. It's not a coincidence that the capacitive reactance of that value cap is 56K, same as the resistor. I would suggest using 10µF @ 150V filter caps though.

I do have that value X1 caps from the previous iteration of this amp.  I will make sure to use the higher watt one.  I did read the amp scrapbook, and I do now remember that part.  I don't have any 10uF 150V on hand right now I don't think, I may have to hunt for some. 

~Phil

[sluckey]

Keep in mind that I specified 10µF bias caps just because that's what the Marshall schematic used. I've always used larger bias caps. Mostly 25µF @ 150V because I had a bagful on hand. Had I built this amp, I'm pretty sure I would have used these nice Sprague Atom 25µF @ 150V caps. But if I had to buy new bias caps I would get 10µF @ 150V.

100V caps will probably be sufficient. But the only way to know for sure is to build the bias supply and put your scope on the 56K resistor to see what the MAX negative peak voltage really is. If the negative peak is less than 100V, then 100V caps will be fine. But if the negative peak is greater than 100V you will need higher voltage caps. If you don't want to make this check just use 150V caps and be safe.

The test circuit on page 6 of my scrapbook only has a max negative peak of 52v so using caps rated for 100V makes sense in that circuit. Your circuit will be a bit different though.

[pompeiisneaks]

Keep in mind that I specified 10µF bias caps just because that's what the Marshall schematic used. I've always used larger bias caps. Mostly 25µF @ 150V because I had a bagful on hand. Had I built this amp, I'm pretty sure I would have used these nice Sprague Atom 25µF @ 150V caps. But if I had to buy new bias caps I would get 10µF @ 150V.

100V caps will probably be sufficient. But the only way to know for sure is to build the bias supply and put your scope on the 56K resistor to see what the MAX negative peak voltage really is. If the negative peak is less than 100V, then 100V caps will be fine. But if the negative peak is greater than 100V you will need higher voltage caps. If you don't want to make this check just use 150V caps and be safe.

The test circuit on page 6 of my scrapbook only has a max negative peak of 52v so using caps rated for 100V makes sense in that circuit. Your circuit will be a bit different though.

excellent, thanks!

~Phil

[pompeiisneaks]

Here's part 4

~Phil

[pompeiisneaks]

Video 5, I got the chassis almost completed, and have more than half of the board done as well:

~Phil

[pompeiisneaks]

Video 6, I finally got some time to work on it again, almost done:

~Phil

[pompeiisneaks]

So I've tried the first power on and it's blowing fuses, after a bunch of troubleshooting, I think I'm beating myself now, because I was completely mistaken.  This isn't a FWBR it's just  FBR.  Here's the screenshot of the original amp (Vox AC100CPH):

So questions.

1. did I kill something?
2. can I just revert to the original schematic and be fine with a FBR?

~Phil

[sluckey]

What is FRB? The power supply in that Vox schematic is a conventional full wave rectifier.

How much time did it take to blow the fuse? This is important. The longer it took to blow the fuse, the more damage could have occurred.

OK, here's what has happened so far. By using a Full Wave Bridge you put twice as much B+ on the filter caps. In this case, that's about 1000VDC. Hopefully the fuse blew before any damage was done. Cross your fingers.

Now, undo that mod and wire the power supply IAW my original schematic. You will not use that separate bias winding on that PT.

[pompeiisneaks]

Full Bridge Rectifier was what I was 'saying' I meant FWR :D

So yeah, it popped almost instantly, in every case, (several times, trying to remove other potential 'causes'.

Luckily on the main first filter cap, the highest voltage I saw on it was about 70VDC, so I'm hoping it kept barely starting to charge and poof.

~Phil

[pompeiisneaks]

Welp

Bad news, I'll be editing the video for it here shortly, but the problem seems to be in the PT.  Yes my boo boo was bad, but honestly, it didn't harm anything, instead once I reset the power rectification correctly, I immediately started up correctly.  As soon as I brought it up for a few seconds though, the 100 ohm fake center tap resistors went up in smoke.  This is the same problem I've had in the past as well on this same amp.  I think the HT and 6.3V windings are somehow crossing over to one another in the PT.  I'm calling this amp a dead beast and moving on.  I don't want to spend another 120$ or so on a new PT to find out the crappy OT is also not working well etc. 

It seems like this chassis from the AC100CPH has been cursed from the onset and it's time to call it quits on this beast.

sigh.

~Phil

[sluckey]

Deja Vu from 2015. Same PT same problem.

Start reading at reply#13. Sound familiar?

     http://el34world.com/Forum/index.php?topic=19596.msg203871#msg203871

[pompeiisneaks]

Yup, I just don't want to buy a power supply after all the money I've throw after this damn thing.  From what I gather the OT's are also quite shit in them.  Here's my final video of the series:

~Phil

[sluckey]

That variac and/or light bulb limiter would have saved you.

[pompeiisneaks]

well it would have saved the two resistors, but not saved the fact that this PT is and has alwasy been garbage :D

But I do get it where I messed up there.  I should have given up a lot earlier with that PT and decided from the onset with the build to either not do it, or pay for a new PT and OT just because. 

Lessons learned across the board :(

~Phil

[pompeiisneaks]

As a final death knell on this, I desoldered the leads on the PT completely and kept the ends covered so they weren't touching anything (This was per a suggestion on FB)  I then put the fuse back in, ran it up slowly on the variac with my ammeter inline.  It got up pretty quickly to 200mA and then suddenly popped the 3A fuse, this with nothing connected.  Proof the PT is toast. :(  Now I guess I may eventually take the time to desolder the board, enough to remove those smoked resistors and then I can restore the board, and eventually I may pay the 120$ for a replacement one, but right now that ain't happening :(. 

~Phil

[pompeiisneaks]

Well, I was able to procure a new PT!  I did some help for someone and they paid me in iron. 

I've got a heyboer, it's also got a separate 103V bias winding, so if I was to use that, I'm guessing I'd have to readjust the setup yet again, probably need a different dropper resistor and remove the jumper from the B+ coming in before rectification and tie one half to ground the other into the 220k dropper and get it a bit lower?  since the half of the main winding should be about what 160-180v I don't need to drop quite as much.  So here's my math, tell me if It's right?

I used a voltage divider calculator to use the max resistance of R1 (dropper + inline) and the R2 which would be the bias pot and its' lower resistor, which comes to about 75k, (25k pot plus 47k)

If I say it's going to be about 50 VDC coming in from the 108 half rectified.  (right?)  then we put in 50 vdc in, wanting about 40vdc out (that's probably max negative I need if my calculations from merlin's el34 / 10 from VDC B+ of around 400VDC = -40DC as the safe starting point)  So in the calculator that give me a dropper of 18.750K. 

So I'd want to go to the next closest, but not sure if 'down' or 'up' is smarter here?  we want generally 'more' negative b ias for safety, but don't want too much or we can't bias warm enough.   is 20k 2W right, or do I want 18k?

Let me know if that's completely wrong, of if I'm even making sense

~Phil

[sluckey]

Show us the spec sheet for that PT. A 103V bias winding suggests this PT was not really meant for 2xEL34s that only need about -40vdc for bias.

[pompeiisneaks]

I don't have 100% positive confirmation, but it's a heyboer and I can give you the writing on the top, it literally looks exactly like what would be the type Doug sells, but the colors are tiny bit off 

Here's the link to the one I think it is:

https://el34world.com/Transformers/files/BRITISH-Plexi-800-100W-PT.pdf

The Numbers on it are:

HTS-6010
9960433

It has two primaries set for use and the remainder are cut shorter and capped.  The long primary leads are orange and red.  (which matches the expected 120V from the above diagram)

The other covered leads there are purple, brown,  blue and what appears to be pinkish/salmon?

The secondary side has:

2 Black (thick) and 1 green (thick) which look like the heaters and center tap (matching the above link)
one white (thin) and one green (thin) which match the 103V winding (bias?)
two blue thin and one yellow thin which seem to be the B+ but the above diagram shows blue, yellow (ct), gray, and these are both clearly blue.

Thus why I'm fairly confident it's the same, but maybe just an older generation, it's got black bells instead of the metallic ones usually on heyboer too.  Maybe a custom wind for someone?

At any rate, i guess I could email heyboer for the exact leads, but that seems to fit pretty closely to me.

~Phil

[sluckey]

First off let me say... the diagram you posted is for a 100W, 4xEL34 amp. Not a 50W, 2xEL34 amp. But you can use it. Let's verify that what you have matches that diagram. First some resistance checks. Connect one of your ohm meter probes to the thin blue wire that you suspect is a HV lead. Now connect the other probe to the thin gray wire and note the resistance. Move the probe to the thin yellow wire and note that you have about half the resistance that you saw between the blue and gray. Now touch your probe to every other wire and make sure there is no resistance between the thin blue wire and any other wire. If all this checks out twist the thin blues, thin gray, and thin yellow together or put a piece of tape around all three. The idea is to keep them grouped together but separate from the other wires.

Now do the same for what we suspect is the filament winding. Connect one probe to a thick black wire and check for resistance to the other thick black and the thick green. These should all read near zero ohms. Now touch the probe to the remaining wires to be sure there is no continuity between the thick black wire and the others. If this checks out, group the two thick blacks and thick green together.

Now check for resistance between the thin green and thin white wire. And make sure there is no continuity to any other wires. Twist together.

Finally, connect one probe to the orange primary wire and be sure there is continuity to the red, blue, and purple wires. And make sure there is no continuity to any other wires. Group these four wires.

Now for a voltage test. KEEP ALL WIRES SEPARATE. Make sure no wires can accidentally touch. Apply 120vac to the red and orange wires. Connect one probe of your AC voltmeter to the thin yellow wire and connect the other probe to one of the thin blue wires. Should be about 178VAC. Move the probe to the thin gray wire and you should also have about 178VAC. Now connect the probes to the thin blue and the thin gray and you should have about 356VAC. If so, the HT winding checks out.

Do the same for the filament winding. Should have 6.3 Between the two blacks and 3.15 between the thick green and either black.

Now check for 103VAC between the thin white and thin green.

Hopefully you now have everything sorted for that PT. Read my next post.

[sluckey]

OK, remember this PT is for a 100W amp and it requires a full wave bridge rectifier. So, tape up the end of that thin yellow wire and tuck it away. It will not be used. Also tape up the thin white and thin green bias wires and tuck them out of the way. You don't need those either. Now since you have already modified your board for a bridge rectifier, just refer back to reply #5 and wire this PT according to that layout. Be sure to check for bias voltage on pin 5 of the output tubes before you actually plug in the tubes! Adjust for max negative voltage. What do you have?

The EL34s will want about -40v and it would be nice if the bias pot will swing between about -35v and -50v.

[pompeiisneaks]

OK, remember this PT is for a 100W amp and it requires a full wave bridge rectifier. So, tape up the end of that thin yellow wire and tuck it away. It will not be used. Also tape up the thin white and thin green bias wires and tuck them out of the way. You don't need those either. Now since you have already modified your board for a bridge rectifier, just refer back to reply #5 and wire this PT according to that layout. Be sure to check for bias voltage on pin 5 of the output tubes before you actually plug in the tubes! Adjust for max negative voltage. What do you have?

The EL34s will want about -40v and it would be nice if the bias pot will swing between about -35v and -50v.

Sorry I wasn't clear, I have 4xEL34 output tubes, it's a 100W amp.  I'm reading the other details now, though, so I'll check back with you in a few after reading/digesting :)

~Phil

[pompeiisneaks]

OK, remember this PT is for a 100W amp and it requires a full wave bridge rectifier. So, tape up the end of that thin yellow wire and tuck it away. It will not be used. Also tape up the thin white and thin green bias wires and tuck them out of the way. You don't need those either. Now since you have already modified your board for a bridge rectifier, just refer back to reply #5 and wire this PT according to that layout. Be sure to check for bias voltage on pin 5 of the output tubes before you actually plug in the tubes! Adjust for max negative voltage. What do you have?

The EL34s will want about -40v and it would be nice if the bias pot will swing between about -35v and -50v.

I actually had to un do that modification, because I had found that I was at like 700VDC with a FWBR.  I just needed the typical FWR instead of bridged.  It's in the later videos I sorted that out. 

So for this one, then I do need to revert back to FWBR Then?  Okay, I had a nightmare time getting the class X cap in there on this board, so I'd rather use the bias windings if at all possible, because I did have to undo that stuff we'd done.  (i.e. the Vox PT that is dead did NOT want FWBR). 

~Phil

[sluckey]

OK, just a slight board mod to use that separate bias winding. The green wire connects to ground and the white wire connects to the board. Notice the ground jumper change on the board! The range resistor in the JCM-800 2203 is 27K so start with that. Be sure you have a good adjustment range of voltage at pin 5 of the output tubes before plugging any tubes in. Change the range resistor up/down to get a good voltage range if necessary.

[pompeiisneaks]

Excellent, thanks so that looks like its the right FWBR setup for the power rectification, and the separate bias winding.  Thanks!

I hope to have some time tomorrow to finish er up!

~Phil

[pompeiisneaks]

Okay wired it all up, PT working as expected, BUT I'm running pretty hot, I'm at 505VDC at the exit of rectification.  I am getting a good -32 up to -48VDC on the bias as well.  That seems pretty hot to me at 505VDC, should I revert back to regular FWB rectification and ground the center tap? 

The amp's working but have a problem.  When switched to the JCM 800 channel it works, but has a bit of hum I'll chase down later, but the plexi side is super duper quiet at max volume on both channels.

I'll also play around wth that but I don't dare leave my caps sitting over 500V on the first two nodes. 

~Phil

[pompeiisneaks]

Okay I didn't have it biased very warm, I checked at it was running at like 17mA or so, and when I dialed it up to 35mA per tube, I got it down to 497VDC.  So it's 'under' the rating of te caps but just barely, doesn't give me a warm fuzzy.  The JCM800 channel has some bang to it, like it.  The Plexi side is super quiet still, not sure why I'm getting so little output.  I'll have to play around with it mroe to figure out why.  At idle I seem to get 60-80mVAC on the JCM800 side but barely 8-15mVAC on the plexi side.

~Phil

[sluckey]

That seems pretty hot to me at 505VDC, should I revert back to regular FWB rectification and ground the center tap? 

YOU CANNOT WILLY-NILLY SWITCH BACK AND FORTH BETWEEN FWB AND CONVENTIONAL RECTIFIER!!!

That PT is meant to be used with a FWB and you ***MUST*** do so. You are right on target with the voltages you have. Bias it up and play it.

[pompeiisneaks]

I wasn't planning on it, just asked the question.  What I'm not sure I understand, is 'if the transformer has a center tap for the B+ side, why couldn't it be used with a FWB"?  What am I not understanding about the transformer being designed for FWBR?  I'm sure it's something stupid I'm missing.  Just that the voltages would come out too low in non FWBR ?

I did bias it up right, it's down in safe territory and I'm only with 2 tubes, I found one of the tubes had a 200ish ohm short to ground in it.  I blew a fuse on first power up (I took it up really slowly on the variac, since I'd had a blowout earlier, wanted to be sure nothing was amiss, and it came out just fine). 

Any thoughts on the low output on the first channel?  They're sharing the same input, so it's not the signal to the grid, somehow the first tube isn't conducting/amplifying as well, or the switching is somehow allowing too much to ground?  Not sure what to look for, I chopsticked around a bit and didn't see any specific problem, but I'm definitely stumped.  I may try switching preamp tubes to see if the weak channel follows the tube or stays with the socket.

~Phil

[sluckey]

I wasn't planning on it, just asked the question.  What I'm not sure I understand, is 'if the transformer has a center tap for the B+ side, why couldn't it be used with a FWB"?  What am I not understanding about the transformer being designed for FWBR?  I'm sure it's something stupid I'm missing.  Just that the voltages would come out too low in non FWBR ?

It's just that simple. That PT puts out 500VDC with a bridge and that's pretty typical for a 100W amp. That same PT would only put out 250VDC with a conventional rectifier. That would make a real wimpy 100W amp.

There is a proper way to use a CT with a FWB. But you don't connect it to ground. Instead you connect it to the junction of two series stacked filter caps to force the voltage to split equally across them. Like this...

     https://el34world.com/charts/Schematics/files/Marshall/Marshall_2203_jmp_master_volume_lead_100watt.gif

Are you sure the Plexi preamp has low output? The 2204 preamp has one more gain stage than the Plexi so it will naturally sound louder. Remember, the 2204 is a big hair screamer. The Plexi is just a crunchy sounding amp.

Can you move text discussions about this amp to a new thread? It's really a drag to wait on those seven videos to load just to read a text message.

[pompeiisneaks]

Okay edited all the work I did to get er working:

~Phil

[pompeiisneaks]

And the last video, I do some geetar pickin!

~Phil

[sluckey]

I think there's something wrong with your recording setup.

[pompeiisneaks]

Why is that?

I gave it a listen before posting, and have another friend that plays and he listened to it and said it sounds good... what's specifically wrong?

~Phil

[pompeiisneaks]

I'm listening to it again, and so far the biggest thing I think that may be problematic is that I record dual audio.  The guitar is miced and into a DAW, and then I have the camera mic .  I then have to manually manage the levels of each recording, and my camera audio seems to be a bit hot, and I didn't do a seamless transition either sometimes so the mic is picking up a tiny bit of some of the amp before I fade it out etc.

Is that what you're referring to?

Any detail more than "I think there's something wrong with your recording setup." would be greatly appreciated.

~Phil

[sluckey]

The sound is very compressed, almost like a limiter is set and will not let the recording get any louder. Then when you "crank" it up it is still the same volume level. Same with your voice. Then at 51 seconds it's like the limiter is switched off and the guitar gets loud for a few seconds (same with your voice). This sudden jump in volume occurs several times through the recording. You don't hear it?

And when you switch to JCM-800 mode it's still the same level but the noise floor and hiss is just as loud as the guitar and your voice.

Let's see if others hear what I'm hearing. I suppose it possible that my old sound card doesn't know what to do with your recording? Anyhow, what I hear on my computer is not pleasing. I listened to several other recordings after hearing your because I was thinking my Klipsch speakers or computer was messing up. Not the case though. Probably not much help.

[pompeiisneaks]

I have two completely separate audio levels I was dealing with/mixing.  My voice going through a shotgun mic on the camera but that clips horribly and has auto gain control to try and limit it that I keep trying to fade in and out.  There were a few points I noticed that I got it a bit off on the fade in/out on my camera mic.  The amp was mic'ed with no compression into a SM57 on one speaker and a sennheiser e609 or e409 I can't recall right now, into my usb audio interface, into cubase.  I then did a mixdown myself.

I think some of what you may have heard was my poor mixing trying to mute out the crap camera audio when the amp was playing, but bring it back in time for my voice.  as for volume levels of the two channels, I'm wondering if something is off with the JMP 2204 channel as at max volume it literally is almost the exact same volume as the plexi channel, and it wasn't whne I had it in the basement/garage.  I may have to peek and see if somethings gone 'off' there.  If you note, in the video I even mentioned it seemed a bit quieter and not as gainy as I thought it should have been and was earlier.

~Phil

[SILVERGUN]

I hear what sluckey is saying and now that you have explained it I think you know what we are hearing.
It's volume levels that are sounding like they are magically fading in and out, and you were doing that in post production mixdown.

I agree with you that the 2204 channel sounds weak. When you dimed it there was no real apparent increase in gain like would be expected. But that was a totally separate issue than the volume levels that were up and down.
There was a point where I said to myself "there it is (at about 3:27)" and then a second later it was gone....probably because you had pulled the other mic level down in your post-mix.