Blackface Princeton build - motorboating at 3/4 Vol

[Voxacthirtee]

Have a Blackface Princeton(no reverb, no trem) build that motorboats at about 3/4 volume and up.
Read all i could on such problems.
The amp works, sounds good until i turn it up past 3/4, and that is actually dependent
on how much the Bass is turned up.

I've checked all the solder joints, reflowed where there might be a question. Changed the coupling caps in the splitter/output stage. Voltages are are in the right range.
Filter caps are 20/20/20, tried a 40 in the first position  just in case 20 wasn't enough, no change.
Tried 10k-100k between the volume pot and 2nd stage grid to curb blocking distortion, no change.

Before i rip apart the EQ section and hang stuff off the pots and lower the coupling cap values in the output is there anything else i should try?

Also-not using Negative feedback

Amp is real quiet, no hum with good star ground setup.

thanks

[HotBluePlates]

Motorboating is a low frequency oscillation, usually because of some common power supply impedance.

The first suspect would be the dropping resistors in the power supply (the ones between each of the filter cap + leads). If one is too small, there may be coupling through the power supply of like-phase signals. That leads to oscillation.

The second suspect would be a poor ground connection. A "good" ground connection would be as close to 0Ω as possible (but your meter likely can only resolve down to a part-Ω before it starts including meter lead resistance in the measurement). A "bad" ground connection may be several-ohms and up. Meter "continuity" may be misleading, because many meters give a continuity beep/tone for resistances as high as 100Ω and above. So you'll need to measure.

All that together means look hard at the power supply, check for proper decoupling resistor values and good ground connections.

[Voxacthirtee]

Thanks,
i increased the resistors between the filter caps, no real difference.
I had dropped them as the power transformer i used doesn't put out quite what an original Princeton does.
I went back through the ground scheme and improved it a bit.
Checked the ground connections to the star ground with an ohm meter. Nil to a fraction of an ohm at most.

changed the .1 caps on the output section to .047, no significant change.
disconnected the .1 in the EQ section(lifted the end connected to the 100k resistor)
Volume full. no motorboating, but i figured that would happen. Replaced it with an .047,
still motorboats, maybe a hair later.

still stumped

[HotBluePlates]

All of the filter caps are new?

The only other cause of motorboating I can think of is if one of the filter caps itself is actually dead or close to 0µF.

The test for that would be bridging a known good filter cap across each of the existing filter caps, one at a time to see where the noise stops.

... I went back through the ground scheme and improved it a bit.
Checked the ground connections to the star ground ...

It's interesting you mentioned a star ground. More problems have been caused by folks "improving" the grounding with a star ground than simply using the original amp's random chassis grounds.

Ideal is something like this:
1.  The filter caps for the output tube plates & screens should have their - terminals connected directly to the power transformer high voltage winding center tap. Hopefully with the output tube cathode grounded at the same point or nearby.
2.  The filter cap - terminals for all the preamp filter caps should be grounded at the same point as the preamp circuits they feed (tube cathode resistor grounds, pot ground, input jacks, etc).
3.  Connect the two grounds (power, preamp) at one point.

The original Fender falls far short of this ideal. But it works because preamp and power tube grounds are separate.

Anyway, motorboating is almost always a power supply issue, and I've outlined the 3 things which can be wrong (not enough decoupling, high resistance (decoupled) grounds, insufficient filtering).

If none of that works, perhaps there is a cold solder joint where a grounded component attaches to its ground wire.

[Willabe]

Have a Blackface Princeton(no reverb, no trem) build that motorboats at about 3/4 volume and up.

When did this start happening? Is this a brand new build?

[eleventeen]

Random idea: Add in 1500 ohm grid stoppers?


What if any is the difference between powering the screens of the 6V6s's in a PR from a no-choke PS node versus having a choke and then separate 470 ohm resistors from the low side of the choke as in Fender 6>>L<<6 amps?


The choke, we know, is a power supply smoothing element. That's if I stand on the rectifier and look out towards the amp. But I imagine it also should serve to block big-swinging plate signal volts from hitting the output tube screen grids. Does that happen?   SOME power output tubes want more tightly regulated screen power so I wonder if that is a desirable effect universally. Comments?

[Voxacthirtee]

Have a Blackface Princeton(no reverb, no trem) build that motorboats at about 3/4 volume and up.

When did this start happening? Is this a brand new build?

yes, a new build with new caps. Usually i can track it down if its a noise or hum, this one i can't.

[Willabe]

So it has always done this from the 1st time you powered it up?

[Voxacthirtee]

Thanks, it's a JJ cap can with 4 connections. I've used them a couple of times without problems. I will try jumpering a new cap across each tap.

After running into ground/hum issues in the past i read a bunch of stuff including the long article on Aiken amps, so i do follow that sort of ground plan now.
This was built into an older mono steel hifi chassis. No way to follow the actual Fender layout, and the star ground seems to mostly work.
1 ground at the earth from the AC cord, close to where the ac comes into the chassis. second ground covers the transformer secondary+cap can+power tube cathode+Neg speaker out+artificial heater center tap, 3rd ground is all preamp tube stuff.
Grounds are on the same physical side/part of the chassis in a straight line from the earth ground.
I could make 2+3 grounds the same point, should i?

In cleaning up the grounds last night i resoldered a bunch of the ground tabs and i'm pretty sure they are ok, will triple check

All of the filter caps are new?

The only other cause of motorboating I can think of is if one of the filter caps itself is actually dead or close to 0µF.

The test for that would be bridging a known good filter cap across each of the existing filter caps, one at a time to see where the noise stops.

... I went back through the ground scheme and improved it a bit.
Checked the ground connections to the star ground ...

It's interesting you mentioned a star ground. More problems have been caused by folks "improving" the grounding with a star ground than simply using the original amp's random chassis grounds.

Ideal is something like this:
1.  The filter caps for the output tube plates & screens should have their - terminals connected directly to the power transformer high voltage winding center tap. Hopefully with the output tube cathode grounded at the same point or nearby.
2.  The filter cap - terminals for all the preamp filter caps should be grounded at the same point as the preamp circuits they feed (tube cathode resistor grounds, pot ground, input jacks, etc).
3.  Connect the two grounds (power, preamp) at one point.

The original Fender falls far short of this ideal. But it works because preamp and power tube grounds are separate.

Anyway, motorboating is almost always a power supply issue, and I've outlined the 3 things which can be wrong (not enough decoupling, high resistance (decoupled) grounds, insufficient filtering).

If none of that works, perhaps there is a cold solder joint where a grounded component attaches to its ground wire.

[Voxacthirtee]

So it has always done this from the 1st time you powered it up?

Yes, and to further clarify and answer eleventeen,
it's actually a Princeton/Tweed deluxish hybrid.

Princeton front end, tweed deluxe power section, sort of.
The outputs sections are close anyways, but i built it as follows.

No negative feedback, no choke, with 1.5k grid1 resistor on 6V6-470ohm/1watt grid2 resistor. Cathode bias - 220ohm+25/25 cap - Slightly over 300v on plates

I've built a couple like this before with the same basic "design" without a problem.

[SILVERGUN]

Are you sharing cathode bypass caps on any/all of the preamp stages?

[Voxacthirtee]

Are you sharing cathode bypass caps on any/all of the preamp stages?

i was just mulling that, you must be a mind reader.....

its the basic blackface preamp using both sides of the first tube, and i did NOT split the cathode.
Used a 25/25+820ohm shared - wondering if i should split?

Second preamp tube uses just half the tube for the splitter/driver. other half(cathode/grid/plate) is grounded.

[sluckey]

its the basic blackface preamp using both sides of the first tube, and i did NOT split the cathode.
Used a 25/25+820ohm shared - wondering if i should split?

Yes, you should ALWAYS use separate cathode components for consecutive stages. Fender shared cathode components in lot's of circuits, but NEVER in consecutive stages. This is likely the cause of your motorboating.

[Voxacthirtee]

its the basic blackface preamp using both sides of the first tube, and i did NOT split the cathode.
Used a 25/25+820ohm shared - wondering if i should split?

Yes, you should ALWAYS use separate cathode components for consecutive stages. Fender shared cathode components in lot's of circuits, but NEVER in consecutive stages. This is likely the cause of your motorboating.

dammit, i should have known that.
I hesitated before i did that, looking at the princeton schematic and thinking they DIDN'T do that because stage 2 on the princeton incorporates the negative feedback, and as i wasn't using neg feedback i'll just use a shared cathode.

I'm a dummy...

will split them and let you all know...many thanks for the help and suggestions

[SILVERGUN]

its the basic blackface preamp using both sides of the first tube, and i did NOT split the cathode.
Used a 25/25+820ohm shared - wondering if i should split?

Yes, you should ALWAYS use separate cathode components for consecutive stages. Fender shared cathode components in lot's of circuits, but NEVER in consecutive stages. This is likely the cause of your motorboating.

Sorry guys, I got busy at work and couldn't keep up...
 
THANKS sluckey, for confirming my thoughts and lending a hand.
 
I remember solving this problem a while back after I had made a similar mistake.
 

[PRR]

> they DIDN'T do that because stage 2 on the princeton incorporates the negative feedback

It's not (just) that.

Two *consecutive* stages with a common cathode causes positive feedback from 2nd stage to 1st stage. Plate swing of 1st stage is buffered by 2nd stage working like a cathode follower, shared to 1st stage, increases effective input swing, often to oscillation.

This matches exactly with your "3/4 up" volume control between stages. Turned down, not enuff PFB to matter; turn up, it oscillates. And at about the same Vol setting no matter what you do with B+ filtering (since the B+ is hardly involved).

It does work with non-consecutive stages. Two input stages. Sometimes a reverb and a recovery stage. In DIY it is almost never worth the brain-pain to work-out what will or will not work. Cheaper to use separate networks than to strain your brain. (In mass production the economics are different.)

It "can" work if the interstage coupling is bass-short (say 50Hz bass-cut) and the cathode network is effective to much-much lower frequency (say 1Hz or about 1,000uFd). That is it works until that oversize cathode cap starts to fade-away, loses uFd. If you don't know this can happen, it is a puzzlement. Again in DIY it is cheaper to put in the extra R-C than to bang your head some years later. (In mass production, different economics, warranty expires, just buy a new amp....)

BTW: Dynaco *used* this cathode coupling to get controlled positive feedback to increase gain. But not directly coupled, a 47K resistor between two 1K cathode resistors. That little bit of leak-back increased raw gain a lot. They then wrapped both stages in heavy negative feedback to get eXactly the desired gain. (So they thought.)

[HotBluePlates]

... to further clarify and answer eleventeen,
it's actually a Princeton/Tweed deluxish hybrid. ...

You probably know this now, but for anyone else reading:
If you ask for help on a "new build blackface _____" but you are not staying 100% true to the original schematic/layout, please say so up front. There are a lot of assumptions we make based on what you say you have to fix.

So this amp doesn't have the same output section I'd assumed from the 1st post, doesn't seem to have the same power rail, has changes to the preamp, etc. Ultimately, it seems one of the (originally) unstated changes is causing the problem.

[Voxacthirtee]

Thanks for all the help, the split cathode seems to have solved the problem.

I do realize i should have been more specific, i just didn't want to get too long winded out of the gate.

I did literally pause and think to myself " now, why again, SHOULDN'T i do a shared cathode here?"
obviously lots of amp use a shared cathode on V1, but of course, they do it when its  dual input stages, not consecutive gain stages. Brain fart on my part.


Thanks again for the help.

[HotBluePlates]

High-Five to Silvergun & Sluckey for figuring out the likely problem!

[Voxacthirtee]

ok, i pushed my luck.
it worked and sounded fine as it was, just not super tweedy.
I had the extra half tube and tried to hook it up as an additional gain stage.

So what i have is essentially the front end of a BF Deluxe amp using the first pre tube.
(dual 68k input resistors, 100k plate resistors, BF tonestack, the split cathodes now, both 1.5k/25) run into the normal output section of a Tweed Deluxe,
using the second pre tube and the 6V6 outputs.

Second pre tube and output section is just like a tweed deluxe.
Only difference is that i am using 470 ohm grid 2 resistors on the 6V6
and the first half of the driver/splitter tube is only biased with a 1.5k resistor, no bypass cap.

I can't seem to bridge the two worlds without problems.

Was using an .02 cap from the plate of V1b into the grid of V2a.
Horrible hum. Or i should say horrible hum after i discharge(?) the cap.
If i don't hit it with the probe checking the voltage i get nothing, then i check it, touching pin 2 on V2,  it discharges(?) and i get hum/signal. While its silent there is no voltage on the cathode of V2a, and  it sees full voltage on the plate of V2a as if the 100k plate resistor is not there. Tried strapping a 220k to ground from pin 2 of V2, no real change. After i hit it with the probe, i get normal voltages on the cathode and plate on V2a.
Of course i tried a different .02 cap from plate of V1B to grid of V2a  , still the same.

Pulling the .02 cap from the grid/pin 2 of V2 hum goes away.
Still get normal "test" noise when i hit plates on V2 with the probe to to test for voltage so its passing signal from that point on still. V2a Plate and cathode voltage returns to normal with the cap pulled off the V2a grid.

Testing for signal/AC at that point, i have nothing on the plate of V1b.
Nary a mv, so it doesn't seem to be carrying hum over to V2.

Do i need a voltage divider? I'm getting all this hum at idle, volume pot down.
I've built stuff with cascading gain stages and have never had this problem.

[HotBluePlates]

... it worked and sounded fine as it was, just not super tweedy. ...

And it won't, because it's a blackface circuit. The tone stack, with its big midrange cut, is what gives the impression of the blackface sound and the cleaner tone.

... So what i have is essentially the front end of a BF Deluxe amp using the first pre tube.
(dual 68k input resistors, 100k plate resistors, BF tonestack, the split cathodes now, both 1.5k/25) run into the normal output section of a Tweed Deluxe,
using the second pre tube and the 6V6 outputs.

Second pre tube and output section is just like a tweed deluxe. ...

If the amp makes sound now (and I'm not sure because of your mention of continued problems), try this:
Disconnect the 6.8kΩ resistor in the tone stack. That is, there is a 6.8kΩ resistor from the Bass pot to ground (if you still have a blackface Deluxe- or Princeton-style tone stack).

Gain and midrange should jump dramatically. Tone controls won't be much use. There's your "tweed sound" in a blackface amp.

If the amp doesn't work now, we'll have to try to sort out what is wrong with your hybrid amp. As with all new builds, there's a 90% chance or better the issue is a wiring problem or implementation problem with the hybrid-amp idea. That's not a knock on you; I make wiring mistakes in my builds, too, but I just sort them out faster than someone with less experience.

It will probably take good hi-res pictures of your actual amp to see what's going on with what you actually have.

[Voxacthirtee]

Thanks HBP.
Ultimately i knew it would not be as Tweedy as i wanted. With the mid notch and gain loss thru the EQ. But i knew i had that extra half tube.
Its not, i suppose that it doesn't sound Tweedy because of the tone, more because of the gain.
I think i can get there with the half tube and no bypass cap. Don't need a gain monster, just something more akin to a tweed deluxe with a 12AX7 in it. Played it full volume for a while the other day and it sounds good, but sounds like a Deluxe at 5 or 6. Just want that non NFB full volume tweed deluxe growl.

I can take a pic, but its "apart" now. I jumpered it back to what i had and it works again.
It worked fine after splitting the cathodes, solving the first problem and i played it off and on yesterday. No problems, only a tiny bit of noise/hum at full volume.

Then tried to incorporate the extra gain stage.
So i'd guess that since it worked fine before i did the added gain stage, V1 and V2b were wired correctly.
So i set about wiring the first half of V2, that had been grounded,  like the gain half of
a Tweed Deluxe splitter/driver.(or driver/splitter)
I ran a 1.5k resistor to ground from pin 3 (cathode) of V2.(did NOT use a bypass cap)
Connected an .02 cap from the plate of V1b to grid/pin 2 of V2a, then added a 100k plate resistor to pin 1 of V2a and ran an .02 from said plate/pin 1, to pin 7 of V2b.
Its built into an old hifi chassis, so using a bunch of real point to point connection with a few barrier strips. No fibre boards.

But it doesn't want to play nice. Again, at idle, with the volume off i get huge hum.
I've seen cascaded gain stages that use voltage dividers, but i don't think i'm pushing anything into the new gain stage at this point. as mentioned, AC millivolts are all but zero at the plate of V1b at idle prior to me adding the gain stage. Yet it seems that connection is whats causing the problems. But i'm using a good new cap. DC is blocked. V1b and V2b work fine without the gain stage added. using a 100k Plate resistor to V2a plate/pin1. Using a .02 cap from V2a Pin  1/Plate to Pin 7/Grid of V2b. Again, Cathode/Pin 3 is a 1.5k to ground, no bypass cap. Don't know what the hell is going on.

I could just pull the 6.8 k and put it on a switch, and not use the extra gain stage, but at this point i'd like to learn why its being a problem so i can tackle it in the future if it happens again.

[Voxacthirtee]

here's a pic. As i mentioned its cobbled into an old hifi chassis. That being said, without the added gain stage
it works fine and is quiet, no hum.
Whats missing from this photo is the .02 cap from pin 6 V1b to pin 2 V2a.

[shooter]

So i set about wiring the first half of V2

How many *gain stages* are on the PS Node.  I've found/learned; 1 is good, 2 not bad, but if I put 3 gains on 1 node things go backward fast

[Voxacthirtee]

I thought about that as well. This would be 3 "gain" stages and the splitter section of V2.
Maybe i'll try to split them, have an extra section on the cap/can

[Voxacthirtee]

that wasn't it. Split the load and now each tube has a 20mf cap, 10k between them, no change.

I still get no signal until i put the probe from the meter on the V2 end of the coupling cap going from Plate of V1b to grid/pin 2 of V2a. Using a grid resistor changes that(tried 220k/1meg)
but still get a loud hum.
What am i missing about coupling stages here?

[shooter]

Using a grid resistor changes that(tried 220k/1meg) but still get a loud hum

Changes what?, you now get audio on the grid, with hum, or just hum?  What kinda hum? 60-120hz?

[Voxacthirtee]

Adding the grid resistor i now get audio on the grid with hum, without having to hit it with the probe until it comes up.
Sounds more like 60 than 120
Cathode and grid resistor grounded at the same point as the rest of the grounds on the tube.
As i mentioned before, using just the other 3 tube sections the amp is quiet and hum free.
(had grounded the cathode/plate/grid of the extra stage).
Works fine from startup, just a little shy on overall gain for my taste.
 
Why i'm getting this much hum adding this section is bugging me.

[HotBluePlates]

... Why i'm getting this much hum adding this section is bugging me.

Without a schematic of the amp as-built (that's 100% accurate to what's in your chassis) and/or detailed photos of the whole inside of the chassis, I don't know how we can help find the problem.

Can you get the amp working with just the blackface-style preamp and the tweed Deluxe-style output section? (EDIT: found my answer, I think)  Meaning, forget the added gain until you have a working platform to modify...

I ask because we have a preamp from one amp, power section from another amp, power supply from who-knows, plus an added gain stage. Tube position references are now not useful for us (V1b of what amp? V2 is what exactly? What is the added gain stage numbered?), so we can't know what grid you're grounding, etc. This also means we don't know what parts are supposed to connect where to sort out if its a build error (wrong part value, wrong wire connection), a layout issue (output tube plate next to an input jack), or simply an implementation error (missing grid reference resistor, too many consecutive stages on one power supply node, etc).

... without the added gain stage it works fine and is quiet, no hum.

Let's try going back to the circuit without the added gain stage. Once there, with the amp working stock, try the tone stack disconnect I mentioned earlier, and see if that works well enough for you.

If not (and we can get a good, complete schematic of the amp at that point), we can likely design an added gain stage or other gain boost.

[Voxacthirtee]

Thanks, i really appreciate anyone who takes the time to read through all this and lend a hand.
I'm probably making it harder to explain/fix than it needs to be because of the way i first built it.

The simplest explanation is that i have the exact first tube of most blackface amps, built almost exactly like the blackface schematic.(use the simple blackface Deluxe as an example). With the following exception

1) Split cathodes using 1.5k/25mf on each cathode

And i am trying to marry it to the second half of a Tweed Deluxe.
(Driver/splitter+power tubes+power supply) The Tweed deluxe part of the build varies in the following ways

1) No bypass cap on the driver tube cathode(pin 3)
2)470ohm resistors on G2 of power tubes
3) Powersupply differs slightly - goes as follows - Using a 5U4GB - the Power tranny has 3Amp/5v tap -

B+/40mf+OP tranny ->1k/1watt-->20mf+Powertubes grid2-->10k--->20mf+driver/splitter tube--->
10k--->20mf+First Preamptube

I can cut and paste the 2 schematics together, but other than the variations above, i have built each part  as in the schematic. I already have it fully functional and quiet without the additional gain stage
When it works fine, i have grounded pins 1+2+3 of V2 and am running an .02 coupling cap from Pin 6 of V1 to pin 7 of V2. Runs fine, no noise, no hum. I have a working platform.

I am familiar with the eq lift, not a fan. I put an eq in here because i want the eq.

If i had built it leaving the second half of V1 empty and had populated the first half of V2 with a gain stage it would have been the same thing- both tubes are used as simple gain stages with 100k plates and 1.5k/25mf cathodes. And that would have also been the proper setup for the first part of the output stage of a Deluxe.

So i can draw it out tonight, but its 2 simple designs and then connecting 2 non highgain gain stages.
I've think i've just done something simply stupid or i'm getting some hum/oscillation from something i'm not familiar with.
I've done cascaded gain stages before without these problems.

Again, thanks for the time and the help

[HotBluePlates]

Thanks for writing that out! I can follow what you outlined there.

Where in the total signal path did you try to add the new gain stage? And am I correct that you were adding the unused 1/2 of V2 when adding this gain stage?

[Voxacthirtee]

yes sir,
unused half of V2 turned into what would be the first half of the Deluxe driver/splitter.
100k plate resistor, 1.5k cathode resistor(no bypass cap). Using  a .02 coupling cap from Pin 6/Plate of V1 to pin 2/Grid on V2, 220k grid resistor to ground, as my attempt to bridge the two worlds.

As per the Tweed Deluxe schematic i am using an .02 coupling cap from Pin 1 of V2 to Pin 7 of V2.
X factor, and something i thought of, is that as its a real point to point build, that .02 is connected directly pin 1 to pin 7. (Pin 1 also has the 100k plate resistor , Pin 7 also has the 1 Meg that goes to the junction of the 1.5k/56k of the Cathode Follower circuit.

Should that .02 perhaps NOT be pin to pin on the tube?

only thing i can think of.

[Voxacthirtee]

tried moving grounds/separating grounds in the pre stage to no success in getting rid of the hum.

oddly, with the bass control at about 1/4 it goes mostly goes away.
Turn the bass all the way down, it comes back. Turn it up more than 1/4 it comes back.