Been Tweaking-Some Observations

[tobykea]

Hi there everyone, I've spent the last few late nights tweaking things in my HRM clone and as well as nearly pulling my hair out, observed some things that seem to be at odds with alot of the comments on this board.

In particular, I've been trying to get my OD tone sounding sweeter after being inspired by a clip Dogears has on here.

I've seen quite a few comments regarding the operation of the ODS sections whereby the first tube clips the top half of the waveform and the second tube clips the bottom half. Balancing the amount of clipping between the top and bottom half is apparently handled by the setting of the ODS 25k trimmer.

From my observations of my amp with a scope, feeding a 1khz sine wave and into a dummy load I can see that the first tube is relatively clean until the first clean volume pot is up to about 12o'clock. Past that point and the stage begins to clip on both top and bottom of the wave. The second stage then clips both top and bottom again but with a more square edged shape depending on the first clean volume and Drive pot settings. Looking at the waveform at the plate of OD2 the top half has a flatish top, rising somewhat until it the comes down heading towards zero volts agin. Going negative then, the bottom half of the wave also has a square bottom but more flat than the top.

Can anyone confirm that these waveforms I'm seeing are typical of this amp?

The reason I ask is that my overdrive sound, although sounding quite nice, has a high end brittleness and a not so smooth sound that I could not detect in the clips from Dogears. I'm just wondering whether or not I have something really squew-if causing this harshness.

I particularly notice the harshness at high volumes and on the scope can see the output stage(6l6's) clipping. I suppose this would happen with everyones amp when forcing the output stage into clipping but I thought I'd sound you guys out.

I wonder where Dogears had his first volume pot typically set in some of those clips because i found that this setting is very integral to the sound then coming out of the ODS section. I can't turn my first clean volume pot past 12o'clock when in Clean mode as the amp starts to break up.

The other observation I made with my amp is that measuring my B+ under full load(just before onset of output stage clipping) dropps down to about 422V. At no load i'm sitting at 455V. Is this normal?

At no load my voltages are:

CL1 Plate 185V
CL2 Plate 195V
OD1 Plate 205V
OD2 Plate 210V
PI 310V
6l6GC 455V.

Originally I had power problems where I was getting about 35Watts out and still i'm only measuring this under full load. Has anyone made power readings of their amps using a scope and had similiar results.

I'm really thinking about starting this thing over again from scratch as it's realy frustrating but i'll hang in there for now and see if I can get some advice.

The networ feeding the OD section is the 470k/47pF into 200K, into the 25K pot to ground via 4k7, wiper of pot through 68k to grid of OD1.

Anyone got any suggestions as to whether my OD section is behaving correctly and whether or not my power supply is behaving correctly?

Cheers,

Toby

[odourboy]

Hi there everyone, I've spent the last few late nights tweaking things in my HRM clone and as well as nearly pulling my hair out.

Welcome to my world as well Toby!

In particular, I've been trying to get my OD tone sounding sweeter after being inspired by a clip Dogears has on here.

Damn that Dogears!!

From my observations of my amp with a scope, feeding a 1khz sine wave and into a dummy load I can see that the first tube is relatively clean until the first clean volume pot is up to about 12o'clock. Past that point and the stage begins to clip on both top and bottom of the wave. The second stage then clips both top and bottom again but with a more square edged shape depending on the first clean volume and Drive pot settings. Looking at the waveform at the plate of OD2 the top half has a flatish top, rising somewhat until it the comes down heading towards zero volts agin. Going negative then, the bottom half of the wave also has a square bottom but more flat than the top.

Can anyone confirm that these waveforms I'm seeing are typical of this amp?

I'm planning some more work on my HRM (replacing the plate resistors with RN65s). If nobody else get's back to you on this in the next couple of days, I'll let you know how my OD section is behaving on the scope.

I particularly notice the harshness at high volumes and on the scope can see the output stage(6l6's) clipping. I suppose this would happen with everyones amp when forcing the output stage into clipping but I thought I'd sound you guys out.

I think this is pretty much expected since the NFB function breaks down when the output tubes start to clip - so the onset of distortion should be rather abrupt. It's my understanding that it's the sweetness you crave is from the preamp distortion, not the power section.

I wonder where Dogears had his first volume pot typically set in some of those clips because i found that this setting is very integral to the sound then coming out of the ODS section. I can't turn my first clean volume pot past 12o'clock when in Clean mode as the amp starts to break up.

Interesting - I can't get any serious breakup from my clean section regardless of volume setting.

Originally I had power problems where I was getting about 35Watts out and still i'm only measuring this under full load. Has anyone made power readings of their amps using a scope and had similiar results.

That's about what I'm getting out of a pair of 6L6GCs (actually, I'm about 40 watts RMS but my B+ voltage is 480)

OB

[llemtt]

Hi Toby

Consider this: it's a completely different affair to design and build a tube amp from scratch compared to tweaking an amp engineered by an experienced builder (think Andy Fuchs for instance): all the obscure work as already been carried out, results are almost garanteed.

I have already completed three "almost" identical amps and I am getting into the fourth but I can't still let them sound identical.

I tell you this because I see too many on this board putting together all the correct parts and 3 seconds later posting all kind of strange problems (too harsh, lack middles, too bassy, no headroom,...) and mainly they are related to the OD sound.

If properly engineered all the reference schematics on this board sound very good from scratch and a little tweaking is only required to reach perfection as per each one's personal taste. The HRM probably is the hardest to build, not a good first choice, but if you like his more "modern" sound it gives you lot of rewards.

Inside a tube amp "signal" voltages range from 1v pk-pk at input, 60-120v pk-pk in typical sections and 400v pk-pk on power tubes, and currents can reach tens of mA; the OD signal has an high-freq content you can't find easily elsewhere; a 1-2W (not carbon comp) resistor physically is something like a 10 turn 3mm coil: a transmitting antenna! where distances between components are an inch at best!! Strange things are to expected..

Yes from what you say your power supply and OD section is apparently operating properly.

Maybe it could be a good idea to post very detailed pics of your build.

my .02
Teo

[Normster]

I know I've hinted at this before, but only dogears sounds like dogears. I rebuilt my first two Dumble-style amps completely trying to get THAT sound. Fortunately, my friend Greg let me borrow a Fuchs-modded Bassman that had been tweaked by dogears and guess what...it sounded almost the same as my amp!

[dogears]

Hey guys!

I always set my clean volume at 1-2pm. Always past noon. I also have the 25K trimmer cranked just about full on

Hope this helps??

HRM likes EL34/KT77 style tubes as well. Think Bluesmaster PI/Presence values.....

Thanks for the positive comments....

Fwiw, speakers and guitar pups make a huge difference.

[llemtt]

..speakers and guitar pups make a huge difference.

player makes difference too! you surely are on a different league

cheers
Teo

[ayan]

From my observations of my amp with a scope, feeding a 1khz sine wave and into a dummy load I can see that the first tube is relatively clean until the first clean volume pot is up to about 12o'clock. Past that point and the stage begins to clip on both top and bottom of the wave.
Cheers,

Toby

Toby, something isn't right. A self-biased 12AX7 stage (connected as a common cathode stage) will always clip the bottom of the input waveform (i.e., the top half of the output since the stage is inverting). The reason for that is this: for the tube to conduct, the grid has to allow for the flow of electrons form the cathode to the plate. Now, when the tube is at idle, the grid is at ground voltage and the cathode is somewhere between 1 and 2 VDC in a typical application. As the grid swings positive when an input signal is applied, the higher potential attracts more and more electrons and so the stage conducts freely until the plate voltage gets too close to the power supply rails, and only then do you begin to clip on the positive cycle.

On the negative cycle, however, when the grid starts to swing negative it won't take long before the tube is shut off. At some point the grid voltage will be too negative with respect to the cathode voltage such that the electrons will not leave the cathode and therefore will never reach the plate. That will show as clipping on the scope.

So, regardless of everything else, I cannot conceive of any situation in which you see "top" and "bottom" clipping the way I understand you observed based upon your description, that is, occurring at the same time, at the same stage and in similar fashion. One thing you haven't reported is the cathode voltages and without that, there is no way to tell wheteher the tube will be working fine or not. As already mentioned, should be a couple of volts or so, positive -- and your grid should be at 0 VDC.

Good luck,

Gil

[Ears]

Gil wrote

> The reason for that is this: for the tube to conduct, the grid has to allow for the flow of electrons form the cathode to the plate. Now, when the tube is at idle, the grid is at ground voltage and the cathode is somewhere between 1 and 2 VDC in a typical application. As the grid swings positive when an input signal is applied, the higher potential attracts more and more electrons and so the stage conducts freely until the plate voltage gets too close to the power supply rails, and only then do you begin to clip on the positive cycle.

Sorry, but that doesn't sound at all right to me. As grid swings +ve the tube will conduct as per chosen load line, when the grid is slightly +ve wrt cathode grid conduction will occur. This may be at a plate current well below that which occurs when the plate voltage approachs "the supply rails" (Vp = 0 volts). Distortion is then highly likely to occur and is now dependant on the grid cct's impedance and on it's ability to deliver power into the tube. You get distortion at both extremes, it is just of a different makeup.

Hey Toby, hope to see you over week end.

[Zippy]

tobykea wrote:
I wonder where Dogears had his first volume pot typically set in some of those clips because i found that this setting is very integral to the sound then coming out of the ODS section. I can't turn my first clean volume pot past 12o'clock when in Clean mode as the amp starts to break up.

odourboy responded:
Interesting - I can't get any serious breakup from my clean section regardless of volume setting.

Then dogears reveals:
I always set my clean volume at 1-2pm. Always past noon. I also have the 25K trimmer cranked just about full on

Fwiw, speakers and guitar pups make a huge difference.

**********************************************************

As dogears notes, the pickups are a governing factor in how these amps respond. When discussing breakup points with respect to settings, it would be useful to know what pickups (brand/model/position) you are running.

Of course, we may not agree on what constitutes the onset of breakup either.

[dogears]

Pot tapers matter too... I use Alpha 1 Meg Audio. 15% taper.

My pups are hums. At the moment, I use Dimarzio Virtual Hot PAF. They are a little hotter than PAF.

Toby, shoot me a PM please.

[ayan]

Gil wrote

> The reason for that is this: for the tube to conduct, the grid has to allow for the flow of electrons form the cathode to the plate. Now, when the tube is at idle, the grid is at ground voltage and the cathode is somewhere between 1 and 2 VDC in a typical application. As the grid swings positive when an input signal is applied, the higher potential attracts more and more electrons and so the stage conducts freely until the plate voltage gets too close to the power supply rails, and only then do you begin to clip on the positive cycle.

Sorry, but that doesn't sound at all right to me. As grid swings +ve the tube will conduct as per chosen load line, when the grid is slightly +ve wrt cathode grid conduction will occur. This may be at a plate current well below that which occurs when the plate voltage approachs "the supply rails" (Vp = 0 volts). Distortion is then highly likely to occur and is now dependant on the grid cct's impedance and on it's ability to deliver power into the tube. You get distortion at both extremes, it is just of a different makeup.

Hey Toby, hope to see you over week end.

I am not sure why you don't agree with what I say, since I believe you quoted half of my story... On one half of the cycle, if the signal is large enough, you will clip when you hit the supply rails. On the other half, you will clip because the tube will shut off. The latter is much more musical than the former and will occur beginning in V1B. You may get the former in subsequent stages is the signal levels are large enough.

Gil

[Ears]

>I am not sure why you don't agree with what I say, since I believe you quoted half of my story... :) On one half of the cycle, if the signal is large enough, you will clip when you hit the supply rails. On the other half, you will clip because the tube will shut off. The latter is much more musical than the former and will occur beginning in V1B. You may get the former in subsequent stages is the signal levels are large enough.

Gil[/quote]

If that is the case then you are correct, and I really quoted the wrong part of your post, it was this that I sought to clarify:

> So, regardless of everything else, I cannot conceive of any situation in which you see "top" and "bottom" clipping the way I understand you observed based upon your description, that is, occurring at the same time, at the same stage and in similar fashion.

If the tube is biased such that the grid swings from both conduction to plate cutoff conditions (with Q point somewhere in between) then I can't see why the wave can't be clipped both sides, as shown in the file I uploaded. I suspect that in most cases, it happens (deliberately) that the tube is biased so as to clip asymmetrically only, since that results in production of predominantly even order harmonics (see footnote on last page of upload) - as you say this is perceived as more musical. I'm basing my argument on texts rather than observation so I'm happy to be corrected.

Cut-off clipping seems to be 'harder' than that resulting from grid conduction. Actually I'm very interested in what happens in the area of operation between onset of grid conduction and the tube approaching the supply rails as you put it (or when Vp=0).

There seems to be contradictory information available. On one hand we are told that in order to reduce the distortion generated by grid current we ought to seek to reduce source impedance 9in the grid cct) . This is evidenced by cathode followers driving class AB2. On the opposite side you get told (e.g. by Crowhurst) to insert grid stopping resisters INTO the grid ccts to mitigate the effect of grid current. I have yet to sort the two opposite directions out. 8^(

[ayan]

If that is the case then you are correct, and I really quoted the wrong part of your post, it was this that I sought to clarify:

OK, I think I understand what's happening here, I probably should have spelled out things better: the grid going negative will take only a couple of volts to occur, so that is the first clipping you will see. The converse would happen when you hit the supply rails, so either the tube bias is way off, or it will take a much larger signal to make that happen. According to the original post, pretty much symmetrical clipping was reported early on, which is why I said that something must have been wrong, since that is not what I have observed in these circuits with the biases involved and the signal levels involved.

Cheers,

Gil