Hoffman Amplifiers Tube Amplifier Forum
Amp Stuff => Tube Amp Building - Tweaks - Repairs => Topic started by: Matty_V on April 15, 2011, 01:16:27 pm
-
Finished my current build and its doing some funky things I can't figure out with my limited but evergrowing knowledge. Its basically an 18watt with an extra preamp tube. If you strum the guitar ever so slightly it will sound ok, just very quiet. But play it any harder and it gets a million times louder and sounds awful...all crackly and fades out as the strings ring. If you crank it up ALL the way and let her rip it won't cut out...tons of sustain but sounds kinda bad. I've tried bypassing the extra tube to no avail so I'm pretty sure its in the power amp. I tried pulling the PI and jumping the signal to each el84 individually and the crackling and fading out goes away. I've swapped tubes around, put in a new OT, new sockets, no difference. I'm stumped.
Another thing that seems fishy but can't seem to find a cause for. When the power and/or standby is shut off the B+ drops to zero in just a few seconds. When the CT to the OT is disconnected I noticed that the filter caps didn't discharge when shut off. The OT I have now is brand new. I'm using dougs multi cap can btw.
Whats going on here? Bad filter caps? Oscillation? Thanks a bunch.
-
Howdy & welcome to the forum. Hard to say. Post the schematic and some gut shots. Sounds like a miswire to me.
-Richard
-
Its basically an 18watt with an extra preamp tube
Depends on HOW the extra tube is added.......if you simply cascaded 2x more triodes for "more gain" then you may have tons of blocking distortion, etc.
As Buttery said, at least a schematic is needed & the gut shots would be a plus........
and, WELCOME!
G
-
For the added tube: I have the first half as a cathode follower which feeds a tilt tone stack, then into the second half as a recovery...then on into the PI. I can't get any guts shots (its pretty ugly with all the unsoldering and resoldering of components!) but I can put up a schematic here soon. Gotta pull it off my work computer.
Thanks
-
Here's my layout. I haven't drawn up a schematic yet.
-
A couple of things on the layout are draw incorrectly.On V2 one of the cathode resistors is 100k.That can't be correct.And the phase inverter shows the coupling caps and driveline caps going to ground before they go through the 470k resistors.
Get your layout cleaned up a bit first.If you wired it like that it's a wonder it worked at all.
-
I wonder if the 2nd triode in V1 is intended to be a cathode follower?
@ Matty_V, in some stages on your layout, you don't have some of the grids connected to anything (not even grid leak resistors) where you have plates and cathode connected. That would be creating problems (like 'no bias'). I think if you draw a schematic of your layout, some things should become obvious. JM2CW
-
PI looks a bit odd.
Try removing the 470k resistor that I circled red.
The one that I marked green is the traditional way to wire the PI. I guess you copied this part from Stout?
-
I don't see any grids not connected??... I have little notes on the pots where the grids go. First half of V2 is the cathode follower, isn't that supposed to have a 100K on the cathode?
Yes the PI hasn't changed from the original stout setup. I was noticing today that it looked a little different than how I've seen them. I'm still in the dark a little bit about how the longtailed pair works. I'll make those changes and report back!
-
...or add the .01 cap that is in the stout:
-
I took that cap out cause I never understood why there were two caps in series right there. Care to shed some light on that for me?
-
BAM! That fixed it. Now to tackle the more fun problem...WAY TOO GNARLY! Past about 1.5 on the vol is completely unusable. Its like a chainsaw thru a PA. Thanks guys!
-
I think that the .01 cap is there because of that 470k resistor that goes to ground. Without the cap the 470k resistor messes the DC conditions of that PI.
Maybe someone could explain this a little better.
edit: Did you add the .01 cap or remove the 470k resistor?
-
I axed the 470K and rerouted the other cap to the other end of the tail resistor. I'm starting to think something is severely worn on my jmp 2204 cause this little 18 watter is louder.
Any advice as to where I should start as far as taming the beast? Perhaps some voltage dividers in the preamp to bring it down to a manageable level? I'm thinking that recovery stage after the tone stack isn't necessary
-
It's gonna be gnarly with 3 gain stages. You might want to split Rl of either the 2nd or 4th stage into 2 220k resistors and tap in between them. Perhaps temp in a 500k pot to dial it in. Your 3rd stage cathode follower isn't boot strapped. That is probably a source of problems. I see you've tilted the PI ala the Stout.
-
I don't see any grids not connected??... I have little notes on the pots where the grids go.
Oops - right you are - I see them now
-
Like buttery said fix the cathode follower. I didn't notice that earlier.
Here is couple options for the recovery stage:
A: bigger cathode resistor (marshall approach)
B: split the plate resistor (fender approach)
-
I don't know what bootstrapping a cathode follower means or does! I haven't used one before so its new to me. Please educate my noob-ish brain. :worthy1:
I'll try out those suggestions and see what strikes my fancy. You all rock face
-
I don't know what bootstrapping a cathode follower means or does! I haven't used one before so its new to me. Please educate my noob-ish brain. :worthy1:
I'll try out those suggestions and see what strikes my fancy. You all rock face
-
>I don't know what bootstrapping a cathode follower means or does!
For a tube to work, the grid needs to be less positive than the cathode. In a grounded cathode amp (gain stage) this is easy because the cathode is usually sitting at about 1ish volt, or whatever Rcathode time tube current is. We can simply slap a 1M resistor between ground and the grid which allows a very little current to leak through keeping the grid at for all intents and purposes at zero volts DC. When you have a cathode follower, now Rk is much larger so the cathode is at a much higher voltage, rule of thumb is somewhere in the neighborhood of 1/3rd Vsource. Now having the grid all the way down at zero VDC is not really what you want. You want the grid to be a volt or so less than the cathode. The quick fix is to boot strap. Keeping it simple for now, assume the tube is idling at 1mA. 1mA time 1k ohms = 1V, so you insert a 1k resistor between the cathode and the 100k cathode resistor. Now you tap your grid leak resistor between the 1k and the 100k. This puts your grid at 1V less than the cathode. There are many other ways to set the grid bias of a cathode follower such as a voltage divider using 2 resistors between B+ and ground, or setting another triode as a constant current source. Bootstrapping is by far the simplest. Fixing this will probably help your amp a lot.
I still think your amp has 1 too many gain stages. I drew your layout into a schematic so I could more easily visualize the circuit. I then redrew it with several suggestions. From left to right:
1. Add a DPDT switch to switch out the 2nd stage.
2. Boot strapping of the cathode follower.
3. Add a master volume after the 4th stage.
5. DC isolate the PI input from Rl of the previous stage using a .001 cap. (I believe you already did this.)
6. Add a tilt switch at the PI. If you play clean, a tilted PI is not your friend. In over drive it makes for good over-tones.
It's been my experience that cathode biased 6BQ5's tend to sound sizzley when bypassed. You might try lifting that 47uF cap and see how it sounds. I didn't really ponder your tone stack. I didn't draw any of the power supply stuffs. Why.....? I'm lazy. :icon_biggrin:
-
Wow thanks for all that info! Super big help. I was getting the impression I had one too many stages, but I didn't see the point in only using half of a tube. Waste! I didn't know what a tilted PI was but you explained it in another post and I get it now. My amp already has a master vol, just currently unhooked. The tone stack is the "AMZ Presence Control" version 2. Its for a stompbox so those component values will probably have to change. I'll try these suggestions out on the bench tomorrow and hear toneful bliss being created.
I originally wanted to use an ef86 for this but read about their tendency for being microphonic. I still may in the future. BTW, what program did you use to make that schematic? I need to get one. I used CAD for my layout but that wouldn't work so well for schematics.
-
http://www.expresspcb.com/ExpressPCBHtm/Free_cad_software.htm
Express Sch. It's free and easy to use. Download it and install it. Next visit the schematic section here. Find a schematic that looks good to you, save a copy, then build your custom library from that.
You might consider dropping a bottle and moving the tone stack to right behind the volume control. The only reason I added a MV was to diminish gain. Look at the Stout schematic. That thang is a little monster. I've built at least a 1/2 dozen variants of that circuit. Don't worry about microphonics. All tubes are microphonic. When you have a ton -o- gain and you tap the bottle, you'll hear it ring. The solution is simple, don't tap the bottle. Sure, V1 being too close to the speaker in a high gain combo can be a nightmare, but that's a layout mistake in my book. Personally I like a pentode front end. FWIW, I hate leaving an unused triode too, but sometimes less is better. (spoken like a true minimalist)
-Richard
-
You can dampen chassis vibrations in a pre-amp tube by sticking an O ring between the socket and the chassis before you bolt the socket on, and also use little plastic grommets in the socket bolt holes. In addition you can slide a couple more O-rings over the bottle itself. (See the EF86 on the R.H. side of the pic)
-
Hey guys. Wouldn't the easiest way to fix this cathode follower be to DC couple it to the previous stage? Or am I missing something here?
-
DC coupling is a heck of a lot easier. I guess I don't understand how DC coupling works (theme of this post!). I noticed the DC coupled cathode follower when rewiring my 2204 but didn't really know what it was. You're shooting B+ and the AC component onto the grid?
Thanks for the tube suggestions. My next build will definitely feature an ef86. How are the new ones from JJ and such? I remember reading somewhere someone recommended nos telefunken.
-
Wouldn't the easiest way to fix this cathode follower be to DC couple it to the previous stage?
That's what I'd do.
-
Wouldn't the easiest way to fix this cathode follower be to DC couple it to the previous stage?
That's what I'd do.
DOH!
Yeah that IS the ticket. Sometimes I just get trapped in the same routine. Ironically the design I'm playing with this morning is direct coupled.
:embarrassed:
-
DC coupling is a heck of a lot easier. I guess I don't understand how DC coupling works (theme of this post!). I noticed the DC coupled cathode follower when rewiring my 2204 but didn't really know what it was. You're shooting B+ and the AC component onto the grid?
Not B+ (which is a constant DC voltage). Its the varying DC coming off the plate of the previous stage. You configure the grid of the DC-driven stage to be at the same quiescent voltage as the plate of the driving stage, and bias the 2nd stage's cathode accordingly. This usually means you have to elevate the heater(s) to avoid h-k insulation breakdown. (Modern production tubes don't handle high h-k voltages very well). The 'advantage' is that you don't have a cap, so you avoid blocking distortion (in that stage).
-
I'm scratching my head again.
With the volume all the way down I can still get a little bit of sound out of the speaker. I replaced the pot, input jack, checked and tightened grounds. Can't figure it out.
I'm also having issues getting the volume level under control. It doesn't seem right to me that it should go from (almost) off to basically full volume with just a tiny twist of the pot. Then as you rotate it more it just gets more hairy. Is this from having one too many gain stages? I haven't implemented any of the voltage dividing suggestions yet. Something just seems off to me.
Also have some pretty intense hiss that goes away when I pull v2, which stays with the socket when swapping tubes. Still lots to iron out. :laugh:
-
With the volume all the way down I can still get a little bit of sound out of the speaker. I replaced the pot, input jack, checked and tightened grounds. Can't figure it out.
This probably happens because there's only one filter cap for the four gain stages.
I'm also having issues getting the volume level under control. It doesn't seem right to me that it should go from (almost) off to basically full volume with just a tiny twist of the pot. Then as you rotate it more it just gets more hairy. Is this from having one too many gain stages? I haven't implemented any of the voltage dividing suggestions yet. Something just seems off to me.
Sounds normal to me. With low loss tone control and three gain stages your input sensitivity is probably < 1mV.
I suggest that you draw us a schematic.
-
I see. Apparently I went a little too far with my design. I originally was going to use an ab763 style tone stack which would have made the extra stages useful. But going to the current stack it seems I have way too much going on. I must ponder my next move.
I learn so much more by doing things wrong
-
Well just to cap off the thread, I've redone my circuit and I REALLY like how it sounds. I don't think the JMP will be getting much attention for a while. I haven't drawn up the schematic yet but I went with....
Input>Gain stage to dc coupled follower>Amz tilt tone stack>Volume pot>parallel triode>PI>Master>Power amp>Out
Good amount of breakup, very responsive to picking dynamics, cleans up well. I noticed with the cathode bypass cap on the power tubes it did have that sizzle sound to it, so I yanked that. I may try a 12ay7 in the parallel section, just don't have one on hand. My board has a whole bunch of unused turrets on it now.
Thanks for the help. Fun project and I learned a whole ton in the process. I'll see if I have time to draw up the schematic at work tomorrow. Feel free to tear it apart :laugh:
-
Here's the schematic. I just realized that jschem is offered for mac os. Is there anything I could do or add to make it work better? I guess "better" is pretty subjective. I'm totally open to suggestions. I've got an extra hole in the front panel, I was thinking to try out a PI tilt switch. Some NFB could be cool too.
-
Oops forgot to put in the gain pot. Revised schem