16uf / 32uf Filtering Switching Option

[Flames1950]

Ah yes, ghosting. You never think it's going to happen to you....

It happens to me all the time.

About the only amp I don't ghost in my arsenal right now is my '75 Marshall 50-watt Lead, mainly because there just isn't that much low-end frequency content (comparatively.) But that just isn't the sound I'm after anymore, so I'm trying to learn to live with and work around it as much as possible. I refuse to keep cranking the filtering up on my amp builds, because it ends up too far away from the sound I wanted to begin with.

[jtm-45/dr504]

How will it change the sound if I add the other side of the 16x16? I'm loving the sound right now, but the ghosting is annoying as hell once you've noticed it.

[Flames1950]

Honestly, I would get the amp into a band situation and see how bad it really is.

The last band I played in I started out with an old Fender Bassman head and an old Bandmaster head. I never noticed any ghosting issues. I've got one of each again now, and stock the ghosting was horrible -- but I haven't had a band around to see how they sound in that context. I just wonder if the ghosts get lost in that situation or not.

Or, it's just one jumper, it will take not five minutes to add it and see if you like it. There was no really radical change in the tone to my ears. And if it helps all the better.

[jtm-45/dr504]

Good advice. I will see if I can notice it in the band, and if so, then I will try adding the other side. Funny, I almost wired up both sides when I installed the board, but wanted to see how it felt and sounded this way first.

[Flames1950]

Also, if you're looking to eliminate ghosting I'd consider not jumpering the two halves together on that can, but running a wire from the empty side over to the side of the 32x32 can that goes to pin 4 of each power tube socket. That's your screens filter, and the screens is the big culprit for ghosting in most cases. That would be the most effective use of the extra filtering; or wire the extra 16uF to the other side of the 32x32 can which is your first filter in the power supply.
Adding an extra 16uF at the phase inverter (what you'd get by just jumpering the can's sides together) isn't really as effective against ghosting as using it elsewhere.

[jtm-45/dr504]

Cool. Thanks for that advice.

[mah84]

As it stands, the planned switching options are: split/shared V1 cathode, V2 cathode bypass cap, switchable PPIMV, and SS/valve rectifier with automatic bias compensation when you flick the switch.

If you are using a standard JTM45 PT, when you switch to SS rectifier you will be at around 500v on the first two filter caps. I have done this, and used the JTM45/100 power supply with the filter caps in series to handle the higher voltage. It might not be strictly required but if one of those cap cans pops, look out!
I also added an extra 30uF filter in parallel with the first filter stage that switches in with the SS rectification. This puts 60uF on the first stage with the SS rectification and takes care of any ghost notes. At least in the SS recto position.

I also have the switchable cathode cap on V2. It's nice to have on a switch but it seems most people that use it like this switched in all the time,YMMV. But it's easy to wire to a pull switch on the normal volume pot without drilling an extra hole.

I use a "fat" switch on the cathode of V1, instead of the split cathodes. This is just the standard shared 820 ohm cathode resistor with a 1uF (or .68uF) hardwired to it. And a 100uF to 220uF cap, on a switch, that is added in parallel with the 1uF cathode cap. This gives you two choices on each channel and can come close to the jumpered tone(not too bright, not too dark) on either channel without jumpering the inputs.

How do you intend to wire the bias compensation switch? Just adding an additional resistance to the bottom of the bias circuit voltage divider?

I like switches, the're shiny.
mark durham

Hi Mark,
Good to see you posting here. I've enjoyed reading many of your posts over in the JTM45 group board. They're always so informative and i've learned a lot from them, so thanks!

When you say there'll be around 500V on the first two filter caps (yeah, i'm using a standard JTM45 PT), do you mean the first two filter caps as in the two halves of the 32uf can cap, or the double 32uf can cap and the double 16uf can cap (sorry if that seems like a dumb question). It appears most of the 32uf cans commonly used by amp builders are rated to 500V, while the 16uf all tend to be 450V.

Also, with your extra 30uf filter that switches in with the SS recto, do you mean to say that ghosting will pose more of a problem with the higher HT? And what kind of switch do you use to take care of all this switching? (how's that for a sentence?)

That fat switch sounds cool. I can see myself fiddling around endlessly with all the switches and component/wiring combinations once this thing is built and (hopefully) running...

The recto and bias compensation switching I intend to use is like in the pic below (I've posted this a few times now in a couple of different forums -in case it looks familiar). Basically there's a separate bias trimmer for each recto, which you set manually beforehand to provide a good idle wattage for each rectifier. Switching the recto also switches to the relevant bias trimmer for that recto. (Pretty much what you said...)


And yes, they certainly are shiny. Very shiny indeed...

[markd]

When you say there'll be around 500V on the first two filter caps (yeah, i'm using a standard JTM45 PT), do you mean the first two filter caps as in the two halves of the 32uf can cap, or the double 32uf can cap and the double 16uf can cap (sorry if that seems like a dumb question). It appears most of the 32uf cans commonly used by amp builders are rated to 500V, while the 16uf all tend to be 450V.

Just the first two 32uF caps in the one can.
Before I built this amp, I mounted SS diodes across the tube recto socket of an earlier JTM45 build. It was set up with a 32/32-500v capcan for the first two filters(like your amp). I have high wall voltage at my workshop(124v) and the B+ voltage on those caps was about 490v-500v. The caps didn't fail but that is too close to the caps maximum for my taste. YMMV.

Also, with your extra 30uf filter that switches in with the SS recto, do you mean to say that ghosting will pose more of a problem with the higher HT?

No, the ghosting shouldn't be audibly worse with higher B+, or SS rectification, but the SS diodes can handle more filtering than the tube recto, so why not add some filtering. I had the power supply designed like the JTM45/100 with four caps in series/parallel at the first filter. The JTM45/100 has 4-32uF caps in series/parallel for a total of 32uF. I replaced that with 4-60uF caps. Two, in series, for 30uF with the tube rectifier. And two in series that switch in parallel with the first two for 60uF with the SS diodes.
I also used two 60uF caps in series for 30uF on the screens, like a JTM45, instead of the two 32uF caps in series for 16uF on the screens, like a JTM45/100.

And what kind of switch do you use to take care of all this switching?

A DPST. One pole to switch in the SS diodes across the tube rectifier. And the other pole to switch the ground of the extra set of caps.

Here is a crummy hand drawn schematic of my power supply. The amp must be on standby to throw the 45/50watt switch.
If I didn't want to switch the filtering, the 45/50 switch could be incorporated into the standby switch with a DPDT(center off) switch. That might work for you to keep from having to drill an extra hole for your SS/tube recto switch. If your interested in that I can draw you a schematic to illustrate.
markd

[markd]

After looking at your schematic again, I see you are using a SPDT switch to switch between the tube rectifier and the SS diodes. I just use a SPST switch to switch the SS diodes in parallel with the GZ34.
A GZ34 has 200 ohms of resistance and SS diodes have only .7 ohms of resistance. When current is given two paths, the amount of current that takes each path is inversely proportional to the resistance of each path. So in parallel almost all of the current goes through the SS diodes. No need to switch from one to the other. Not that that won't work.
One interesting thing is that because of how stiff a GZ34 is, the B+ voltage, at idle, doesn't rise a great deal with the SS diodes. My B+ goes from around 460v to around 490v, IIRC. But the feel changes quite a bit, IMO, due to the lack of sag with the SS diodes. At maximum input signal the B+, with the GZ34, will drop about 40v. That won't happen with the SS diodes.
If you do go with the SS/tube rectifier switch, and you are using modern tubes, I would recommend increasing the screen resistors to 1K. There was a recent post here that mentioned that increasing the screen resistor value would negatively affect the touch sensitivity of the amp. I don't know about that, but from a purely technical standpoint putting almost 500v on the screens of a KT66(max screen voltage 400v) is asking for it. With the tube rectifier, you get much lower screen voltages, with maximum input signal, due to the drop across the GZ34 and the drop across the screen resistors. With the SS diodes, and only 470 ohm screen resistors, those KT66's are really being stressed. I like how the Valve Arts/Shuguang KT66's sound but I have fried a couple of them. Do you know if the "TungSol" KT66's are any tougher?
markd

[Flames1950]

Hmmm, you make me think I should put 1K screen resistors on my JTM45/100 clone.....but it sounds so nice, and the GT KT66HP's are hanging in there pretty good so far with 530VDC on the plates and not much less on the screens.
I tried the Tung-Sol KT66's in that amp for a day, liked the smoother sound of the GT's better so out the Tung-Sols came. So I couldn't add to any discussion of their longevity or hardiness yet.
I've wondered if the T-S's would sound better in my JTM45 clone, with its substantially lower plate voltages.

[markd]

Hmmm, you make me think I should put 1K screen resistors on my JTM45/100 clone.....but it sounds so nice, and the GT KT66HP's are hanging in there pretty good so far with 530VDC on the plates and not much less on the screens.

Measure the voltage drop across your 470 ohm screen resistors, at idle. If it's 3.5v, or more, you are exceeding the maximum screen dissipation for that tube, even at idle. A prudent person would increase the screen resistor value if that were true. Although, prudence isn't always a virtue when tone is at stake.
IMO, tone and feel come first. If you can afford the consequences and if the amp doesn't actually melt down anything goes.
markd

ps. I have never used the GT KT66's due to some old emotional problems with GT (ha!) but it sounds like I should give them a chance.

[Flames1950]

ps. I have never used the GT KT66's due to some old emotional problems with GT (ha!) but it sounds like I should give them a chance.

Well, George was blowing them out cheap due to how many problems he'd had having to take them back after the sale....I couldn't refuse $40 a quad. I don't appreciate Aspen's marketing BS any more than anyone else, but then the "Tung-Sol"-marketed tubes rub my fur the wrong way too.........

[MacGaden]

If I didn't want to switch the filtering, the 45/50 switch could be incorporated into the standby switch with a DPDT(center off) switch. That might work for you to keep from having to drill an extra hole for your SS/tube recto switch. If your interested in that I can draw you a schematic to illustrate.
markd

I´d like to see that please. Sounds interesting.

[mah84]

Mark, thanks for the replies.

When you say there'll be around 500V on the first two filter caps (yeah, i'm using a standard JTM45 PT), do you mean the first two filter caps as in the two halves of the 32uf can cap, or the double 32uf can cap and the double 16uf can cap (sorry if that seems like a dumb question). It appears most of the 32uf cans commonly used by amp builders are rated to 500V, while the 16uf all tend to be 450V.

Just the first two 32uF caps in the one can.
Before I built this amp, I mounted SS diodes across the tube recto socket of an earlier JTM45 build. It was set up with a 32/32-500v capcan for the first two filters(like your amp). I have high wall voltage at my workshop(124v) and the B+ voltage on those caps was about 490v-500v. The caps didn't fail but that is too close to the caps maximum for my taste. YMMV.

I'm in Australia, so i'll be running 240V into the 240V tap, but I know what you mean, that is cutting it a tad close. Might have to try and source a higher rated cap can, just to be safe.

That's some cool filtering/rectifier switching you're using there, very crafty indeed. Now you've got me thinking of new possibilities, just as I was trying to cut back and simplify things! You say that while the B+ doesn't rise much with the addition of the diodes, the feel of the amp changes quite a bit. Other than the sag factor from the GZ34, do you think the extra filtering has much to do with this change in feel? Regarding the switching diagram, I've seen pics like yours in the past where the diodes are switched into parallel with a tube, but I wasn't sure exactly how much of the current would travel via each pathway -I wanted to be sure the current was travelling either entirely through the GZ34 or entirely through the diodes. But yeah, with those resistances, I guess using a DPDT is kinda unnecessary.

Also, I'll actually be using the Marstran JTM 50 OT (3.5k primary) with EL34's so there shouldn't be any problems with screen voltage using 470ohm screen resistors.