Announcement

**otto p�rt** · 02-01-2010, 01:37 AM

How far away from the preamp circuit are the power tubes and output transformer?

**imo** · 02-01-2010, 02:03 AM

Thanks for the reply.
The preamplifier tubes are to the far right of the chassis, with the ot left middle, and the power tubes slightly to the right of it

**pdf64** · 02-01-2010, 09:27 AM

If a stage is no longer sharing a cathode resistor, then it will need to be upped in order to retain the same operating conditions.
Motorboating is generally down to power supply decoupling - try subbing in a known good cap for each B+ cap in turn, to rule out a bad cap.
If all ok, does the motorboating stop if the phase splitter tube is removed (measure the motorboating signal at the input to the phase splitter)? If so, you may need to add another level of RC decoupling on the phase splitter, if not then an extra stage decoupling on the 1st preamp tube.
High frequency oscillation in an proven circuit (but unproven layout) will probably be due to lead dress. Sounds like your layout has got a lot of voltage gain going on with not much physical seperation between the 1st grid and power tube plates, so even with lead dress as good as possible, there may still be a problem, requiring a circuit mod.
Best practice from the old days was to fit grid stoppers (say 10k-22k) to each pre-amp grid at the socket terminal - even just moving the 68k input mixer resistors from the jack sockets to the socket terminal can be beneficial.

**imo** · 02-02-2010, 04:00 AM

thanks PDF,
I did try subbing in other caps(paralleling them over the orig caps) the cap can was brand new so i didn't think it would be, and indeed i found little or no difference there.
I will try looking at the PI input. I am assuming you are saying on a scope. I don't have a ton of experience with my scope, so i would look for a signal riding on a sine wave, or is there an easier way to do this?
The design already has 5 stages of filtering. i would be willing to add more, but trying to find a way to make the proven design work as is before i start modding it, possibly covering up an error on my part.
I have the initial 68k grid resistor at the grid. I don't currently have grid resistors on V2 as i'm trying to keep it stock, at least for starters.

Could there be something in the grounding scheme? I remember fixing a B15 and it was doing a similar thing(low end oscillation at high volumes) I had to mess with the cap can grounds a bit before i was able to clean it up.
Ian

**imo** · 02-02-2010, 04:53 AM

btw,
just went out to the studio to take a look the amp. The grounds are a bit wonky, I had the 2 filter caps feeding the preamp grounded through a lug and also to the main filter cap ground. I cut those, but it didn't change anything.
I did notice however that i have the speaker out ground connected to the preamp filter cap ground. Could this be causing the oscillation?
I won't be able to drag the amp into my shop until tomorrow evening, but it would be good to hear people's opinions on grounding as it still seems a bit complicated to me
Ian

**pdf64** · 02-02-2010, 09:49 AM

'I did notice however that i have the speaker out ground connected to the preamp filter cap ground. Could this be causing the oscillation?'
As there's no negative feedback circuit, where you ground the speaker circuit shouldn't make any difference.
However, on my '63 AC30 (copied from AC50s I've worked on) I've added screening plates to shield the input jack sockets and OT secondary/terminal block.
Without it, at high settings, these things can oscillate, especially if the gain has been upped a bit, which your's will be, with only 1 channel.

'design already has 5 stages of filtering'
Each pre amp B+ node only has 3 filter stages (ie plate node, screen node, then 3 parallel pre amp nodes) as stock, but as you're only on 1 channel there may be a spare pre amp B+ node (ie for vibrato channel).
I think you might need an additional filter stage in series, like a fender BF arrangement.

'I will try looking at the PI input. I am assuming you are saying on a scope.'
You may be able to measure the motorboating signal with a multimeter set for AC at the phase splitter input.

**imo** · 02-02-2010, 03:53 PM

I was able to see the dc oscillation at the PI plate. When i pulled the tube out it did steady. i could definitely then use some more filtering.
I am actually looking to build a "clean" ac30. Lots of power and headroom, with most of the distortion coming from the EL84's, so i wouldn't mind lessening the preamp gain to start with. That could take care of some of the oscillations as well i would imagine.
It seems that my tone section is being hit too hard in general. The bass seems pretty hefty at 1/3 to 1/2 up, and starts oscillating at 50% so it seems that maybe lessening the gain coming into the tone section would help
I'm going to try messing a bit with the 1st gain stage. Maybe try a 150k or so plate and possibly a slightly bigger cathode r as well. Does this seem like it would help to pull it together?
i will also try a bit more PI filtering
Do these moves seem to be generally in line with getting a cleaner(and hopefully non oscillating build) to start with?
how hard do you push those 84's? It seems that 17w is a lot even for this design.

**pdf64** · 02-02-2010, 04:21 PM

OK, try taking the feed for the pre-amp B+ from the phase splitter B+ node. If the B+ voltages come down too much, reduce the resistor between the screen node and phase splitter node from 22k, try something between 1-10k.
To bring the gain down, try a 12AY7 for the pre-amp.
Or a potential divider on the signal before the phase splitter - try a 1M preset pot
Messing with the plate resistors etc doesn't do a whole lot - I've got 1 channel at 220k, the other at 47k, and there's barely any difference in perceived gain, it's more of a tone thing.

**imo** · 02-02-2010, 11:19 PM

Pdf,
It looks like i found the oscillation, and of course, it was user error. The node feeding V2 was connected to the wrong end of the 10k, thus it was out of circuit. I now can turn it up to ridiculous levels without it motorboating.
Thanks for helping me through that.
Now i'm still feeling like the amp could be cleaner before the powertubes. Would you still suggest a 1M pot before the PI? put it after the 220k series, and before the .047 cap?
I will try the 12AY7 as well. I did that a couple of years ago when building more of a DC30 style amp. It was too gainy, and the AY7 helped to tame that.
Out of curiosity what are your thoughts on using different plate resistors on the early gain stages(V1, etc)? I understand technically what happens to the plate curves, but would be curious to hear your ideas on what it sounds like from experience.
Also, on V1, what is the best way to pick a "more appropriate" cathode resistor for only using 1/2 of the triode? would you graph it out on a tube chart or just plug and play? Don't mean to sound like a total noob, but the designers of the great amps seemed to follow a few paths, and most modern schematics seem to follow those pretty closely as well

thanks for your info!
Ian

**imo** · 02-03-2010, 05:37 AM

I tried the 1m(actually 2m) pot, and its pretty cool. Definitely makes the range of volume more usable, seems to subtly change the tone as well.
An improvement indeed.
Right now i have a 5751 in the 1st spot as well. It seems a bit more mellow on the front end in a good way.
The treble is a bit shrill, even for an AC30, but its getting a lot more usable.
Need to tweak it some more but definitely getting closer

**imo** · 02-04-2010, 04:46 AM

Maybe i should readjust the topic? I'm curious to hear anyone's opinion about a few different areas of AC30 type builds. I'm a pretty big fan of the design, and trying to actually understand the circuit better.
things i'm considering:
- changing cathode resistor. I am currently trying a 100r, which has the tubes running within the 12w max diss. It sounds pretty good and i'm wondering if it would be a better move than the higher current of a 50r or so (for me this was almost 17w dissipation!) i would imagine this could severely limit tube life and make for an unreliable touring amp
-upping screen resistors- Lowering the screen voltage by a few volts helps with crossover distortion? Pros and cons?
-possibly changing tone stack. something bigger for the treble cap would allow it to focus a little lower in the treble pot. Also possibly replacing the 10k mid resistor with a mid pot. Currently the amp is a little thick in the low mids and i'm thinking it could help with this..

Anyone have any thoughts on any of this stuff? Learning is nice<

**pdf64** · 02-04-2010, 09:28 AM

I would be concerned that a cathode resistor that high (100ohms) would be limiting the power output and causing crossover distortion at high output levels, because the cathode voltage will rise significantly.
You can see this for yourself as you’ve got a scope – view a sinewave across a dummy load as you increase the signal level, you’ll be able to see indentations appear on the rising and falling parts of the wave.
This is probably why cathode biased amps tend to run at, or close to, class A – the closer to class B the bias is set, the more the bias voltage will rise with increasing signal level.
By using a high B+ power transformer you’ve exacerbated the el84 plate dissipation problem that the original design has, which tends to run them at or over spec.
Your best bet would be to use the military sovtek el84 (some suppliers call them el84M), which have a higher plate dissipation rating, and experiment with cathode resistors.
I’d guess 68-82ohms might be a suitable compromise, but your B+ level is above what I’m used to.
I prefer to use a separate cathode resistor (and bypass cap) for each el84, which allows them to self balance themselves, 240-270ohms being values I’m happy with at my B+ levels (equivalent to a shared resistor of 60-68ohms).
Re screen resistors, as I understand it, increasing them will make crossover distortion more likely. 100ohms does seem to be too low for current production el84s though my old mullards seem fine.
See
JJ_EH_tube_tests
for a study on this topic.
Re your over bright tone, try lifting the bright cap (100pF) bypassing the volume control.
However, what I’ve ended up doing is using the Marshall Plexi arrangement for the basis of the pre-amp / tone controls - it just works better, for me.

**beefwhiff** · 02-04-2010, 02:11 PM

For more headroom, you'll want a lower cathode R. Stick w/ a shared 50R and somehow reduce your B+ by clamping. 7189's instead of EL84s will give you a couple more watts Pa if needed.
Thick low mids? Most would say the stock TopBoost is too shrill. This is why many opt for the normal channel or "normal" and "treble" model configs. Also, the TB will give some grind. If you're looking for headroom, keep it simple w/ single triode stage into PI, IMO.

**imo** · 02-04-2010, 03:48 PM

Thanks PDF. I put the amp on the bench and you're right, the higher cathode resistor is creating addtional crossover at higher input levels. I'll go back to something around 50. Any suggestions for the most rugged way to drop the voltage a few(20) volts on the HV? I imagine putting some resistance before the centertap of the OT would change the way the amp sounded somewhat. I've heard mixed reviews about using zeners. I would worry about reliability.
This will be my main touring amp and will be run hot and long. I am building this more like a marshall physically, so the tubes will have more ventilation, but still trying to bring them to a place where i can get a good deal of reliability.
Currently i have a set of JJ's in there. that sound pretty good. They have held up pretty well in my other builds. My DC30 style amp was played for months at a time with no problems with similar wattage through the tubes..
In terms of the amp being thick in the lower mids, it is a bit, and can get a bit shrill as well. There is a fine line in that design, to my ear, where the low end gets to flubby. Part of it is the sound of the amp, at least the ones i've heard, but it does seem that there are some low mids that could disappear while keeping the low end, thus cleaning it up a bit. I'm also working on getting rid of some of the shrillness. I like it to a degree, but its just too much through the run of the treble pot. I may try the plexi style tone stack to see if that makes a difference. I have already upped the 'bright' cap to 250p, may try removing that as well..

I may try bypassing the 2nd tube just to hear it, straight from V1 to the PI. Definitely worth a go. I am digging the sound. Its just too much coming into the Power tubes without dumping some gain before the PI currently
Ian

Announcement

AC30 style build oscillation in bass and cut controls

AC30 style build oscillation in bass and cut controls

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment