Cleaner sound

[Tdale]

I have a Matchless DC30 clone which I think distorts too early.

What would you guys try, to make it distort less/at higher volume?

I'm no expert on reading tube charts etc. But is there a simple way to measure the preamp tube voltages, to find out if the bias point is where it should be, to allow for the most headroom?

Tommy

[Darkbluemurder]

Since the Matchless DC 30 is modelled after the VOX AC 30, early distortion is part of the design. If you really want loud and clean then a DC 30 clone is not the best amp for this purpose. You would be better off with a Twin Reverb or a Hiwatt clone for loud and clean.

I guess you have tried the options that the design gives you. Best to use the top boost channel (the one with the bass and treble controls) rather than the EF86 channel. You may have to turn down the guitar, too.

If you really want to get more clean headroom out of the amp I would try fixed bias on the power tubes. The DC 30 is cathode biased. You can derive the negative voltage from the high voltage line with a diode, a cap and some resistors. I don't have a schem ready but you can do a search in the Trainwreck forum for "bias" which should give you an idea where to start.

Good luck!

[Tdale]

OK. I'm aware that it's not the best amp for loud and clean sound.

I was more thinking about small tweaks to give it just a little more headroom. How do I calculate the "perfect" cathode resistor for the preamp tubes for instance? By perfect, I mean the bias point which gives the cleanest sound.

There is also a question about gain, that I don't seem to be able to understand:

A 12ax7 has a gain factor of about 100. That means that the output voltage is roughly 100 times higher than the input. But can this go on and on when you put several tubes in series? Or is there a limit of how high input voltage there can be at the grid, where the output voltage won't be 100 times higher than the input?

Tommy

[ampgeek]

Hey Tommy,
Here is a "mirror image" thread that you might find interesting:

https://tubeamparchive.com/viewtopic.php?t=3219

Consider small tweaks to plate resistors and cathode resistors (down and up respectively) to start. Since you already have a well functioning amp, going back to the beginning and attempting to calculate "optimum" values might not get you there any faster. There is just so much dynamic interplay in all of this it is likely you will be lead down a rabbit hole.

Mind you, doing it on paper a few times is invaluable for gaining insight into how all the "pieces" fit together. It might all make a little more sense to you after having actually seen/heard the practical implications.

Also, DBM is correct (based on my limited experience with his suggestion), switching over to a fixed bias power tube plan will give you the biggest bang for your buck. However, you will now have a significantly different amp.

Just my thoughts,
Dave O.

[dobbhill]

I may be misinformed (or wrong), but my understanding is that EL84's don't produce any more power in fixed bias than they do in cathode bias at normal voltages.
D

[ampgeek]

Hey D,
Technically, you are absolutely correct. The max power is typically limited by plate and/or (for EL-84 and other pentodes) dissipation.

However, the normal convention for an amplifier's power rating is maximum "undistorted" power. Generally, at the point when the control grid voltage equals cathode voltage. As the control grid starts going positive, you may squeak out a bit more power but, at some point beyond, it becomes un-pleasant/too distorted.

The nature of cathode biasing allows the grid to "sneak" into that nether world of causing distortion during a portion of the grid's AC voltage swing at "max power".

Fixed biasing is like a wall. You can run the grid right up to bitter edge of causing distortion at "max power" without breaking through. Thus...a cleaner/lower distortion signal at that "max power" point.

Of course, these are generalities and there are other things going on that may effect this.

Dave O.

[Ears]

OK. I'm aware that it's not the best amp for loud and clean sound.

I was more thinking about small tweaks to give it just a little more headroom. How do I calculate the "perfect" cathode resistor for the preamp tubes for instance? By perfect, I mean the bias point which gives the cleanest sound.

There is also a question about gain, that I don't seem to be able to understand:

A 12ax7 has a gain factor of about 100. That means that the output voltage is roughly 100 times higher than the input. But can this go on and on when you put several tubes in series? Or is there a limit of how high input voltage there can be at the grid, where the output voltage won't be 100 times higher than the input?

Tommy

Here is a take on a graphical method to find optimum 0perating point. Its iterative and will quickly give good visual idea of where you should be.

You need plate characteristic curves, and need to calculate DC and AC (signal) load lines.
DC load line easy, its slope plotted on the plate curves is just = -1/Ra (where Ra is anode resistor (we ignore effect of cathode resistance if its very much smaller)). Drawn from x axis at supply voltage (Ip = 0) to point where it intersects Ip axis.

Ac load line (L-L) uses Ra in parallel with following load of following stage. Easy if that's just RC coupled to following grid, harder if includes a tone stack (you can reduce the network for each case: low, mid and high freq bands). Its slope is -Ra//R(load)

Choose a bias voltage. Where its curve intersects the Dc L-L is your operation point. Draw AC L-L so that it intersects at that point (ensure it has correct slope).

The tube will swing along the line. Visually you can quickly get an idea of input signal swing limits and if the two sections of the L-L either side of the Op point (between Vg =0 and Ia = 0, outside that the tube is in grid conduction or cut off) are equal. Equal means, at that op point and given load, the tube has max headroom, and least distortion.

Choose cathode resistance so that chosen Vg = Ip (at that op point) x Rk.

Try a different grid biases (i.e. change Rk). Change load, Ra and start over if you wish.

The method is approximate.

Ref attachedRDH4, Terman

[dehughes]

I swapped a 12AU7 in for the stock 12AX7 in V1 of my Top Hat Club Royale (AC30-ish preamp) and it provided a TON more headroom....FWIW...

[PRR]

> A 12ax7 has a gain factor of about 100. That means that the output voltage is roughly 100 times higher than the input.

Define "roughly".

The tube has gain, everything else sucks gain. Even the resistor you need to feed power to the tube without shorting its output.

At best you get 50%-70% of the theoretical amplification factor.

12AX7 has a lot of gain and gets sucked down more than 12AU7. There's a 5:1 ratio of amp factors, but in actual use the gains may be only 3:1 different.

> But can this go on and on when you put several tubes in series?

Yes. One stage 12AX7 is gain near 50. Two stages is gain of 2,500. Three stages is 125,000. Four stages is 6 million something.

You never see a gain path like this in audio, but FM radio IF strips sometimes have cascade gain near a million.

> is there a limit of how high input voltage there can be at the grid, where the output voltage won't be 100 times higher than the input?

Three stages, put in 1 Volt signal, get out 125,000 Volts signal......

But wait! We only have 300V supply! How can we get 125,000 signal?

We can't.

In fact the peak low-distortion audio output of a standard triode stage is maybe 20% of supply. Or 60V max signal for a 300V supply.

So the maximum input to a 12AX7 is maybe 60V/50= 1.2V.

For the stage before it to deliver that 1.2V, it needs an input of like 1.2V/50= 0.024V. If we have no other losses, the maximum clean input to a 2-stage 12AX7 amp is 0.024V or 24mV. And that's a fairly soft stroke on an electric guitar pickup.

But we have losses in tone controls. We can't actually boost anything; we cut everything and add-back the stuff we want to "boost". Most strong tone controls have loss of 20 or 30. We tend to add other frills, which need mixers, which have loss of 2 or 20 depending on function. So it is fairly normal to want three 12AX7 stages between guitar and big bottle grids, or get by with two stages on low-frill small power bottle amps.

BTW: the tube hiss is about 5 microVolts. So if you did stack four stages of 12AX7 and zero added losses, the self-hiss would drive big 6L6 to full output.

> my understanding is that EL84's don't produce any more power in fixed bias than they do in cathode bias at normal voltages.

The whole idea of EL84 is "cheap". Self-bias is cheap. The EL84 is specced for a fairly high plate dissipation and a fairly low plate and G2 voltage. If you stay inside the Ratings, it will make good power in self-bias and not really any more test-bench power in fixed-bias (will idle cooler).

The 7189 is an EL84 with a 400V rating. If you leave it self-bias and crank to 400V, it'll melt at idle. But if you fix-bias, you can control the idle heat yet still reach higher power output. Like 17W self-bias, 24W fix-bias.

And in fact most EL84s will survive 7189 operating conditions. Especially today after decades of cheap EL84 amps which exceeded the original Ratings; today's tube makers don't dare make an EL84 as cheap as the original ratings claimed.

But "conversion" from self- to fix-bias is really a major re-design. And quite often not worth the trouble and risk.

===================================

> I think distorts too early. ... is there a simple way to measure the preamp tube voltages, to find out if the bias point is where it should be, to allow for the most headroom?

I'm shocked nobody has told you the obvious possibility: the amp may be sick.

Tube amps can be AWFUL sick and still "work".

It would be good to compare against a factory voltage chart and against an identical known-good amp. (If the other amp distorts the same way, that may be "how it is".)

But in ignorance, check this.

At each stage, identify the plate and cathode resistors. One end of the plate resistor is at plate voltage, the other end is at the B+ for that stage.

Measure both ends. The plate should sit at 1/2 to 2/3rd of the B+ voltage. That's not always true, but anything very different should be investigated.

160V on plate, 250V B+, is reasonable and probably not sick.

230V on plate and 250V on B+ is awful funny.

There should be cathode voltage. In principle you can compare voltages and values on the two resistors (plate and cathode) and prove that they flow the same current. But the most likely faults are a shorted cathode cap (zero cathode voltage!) or a drifted-high/open cathode resistor (more than couple volts on cathode).

There are goofy stages which break these rules. Cathode followers are upside down. Cathodynes split half and half. Longtail splitters stand on a long tail. Marshall likes a stage with a huge cathode resistor to get low gain and clean-to-a-point performance.

If the volt-amp stages seem reasonable, replace the power bottles. Don't argue, just do it. Especially on these over-hot self-bias EL84 amps. If it makes a difference, the old bottles were worn out.

Don't forget to try another speaker. I've been amazed at the funny sounds speakers can make. Usually clearly a mechanical problem, but sometimes very strange noises that sound like amp faults.

And if this is a Pentode Input, you probably MUST turn your guitar down or it'll be bent at the first stage, and it don't matter how you set or set-up the rest of the amp.

[titanicslim]

Right. And don't forget to try that 12Au7/12AT7.

[Tdale]

Increasing the cathode resistor of a preamp tube, lowers the bias point. But will that give me cleaner sound, or will it distort earlier in the lower half of the cycle, giving me more distortion that if I bias it in the middle of the curve?

What about the output tubes in a push/pull amp? Will lowering the bias point give less distortion? Will it also give me less power?

Tommy

[Ears]

> Increasing the cathode resistor of a preamp tube, lowers the bias point. But will that give me cleaner sound,

To make a comparison it is needed to know where the bias point was originally, and how big the input swing is.

> or will it distort earlier in the lower half of the cycle, giving me more distortion that if I bias it in the middle of the curve?

I would say yes to that. But going the other way will result in similar result from a different cause.

> What about the output tubes in a push/pull amp? Will lowering the bias point give less distortion? Will it also give me less power?

There seems to be a discrepency in what you mean by "lowering" here than in the first questions. If in class AB you make the bias more negative you will approach Class B operation, higher efficiency and greater power.
The distortion spectra of in class B is different from class A. Crossover distortion becomes an issue in class B and beyond.

Here's some more from Terman on it (not as involved as RDH4)

[David Root]

Ears, where can I find a download of the entire Terman book, and what's the title? It looks pretty solid, lots of math but practical too.

[Ears]

Ears, where can I find a download of the entire Terman book, and what's the title? It looks pretty solid, lots of math but practical too.

Frederick Emmons Terman was professor of Electrical Engineeering at Stanford University 1930s-60s?.
He wrote three useful titles

1) Radio Engineering 1932, 1937 (2nd Ed considered best)
2) Radio Engineer's Handbook 1943
3) Electronic and Radio Engineering 1955

I scanned four chapters of a borrowed copy of 3). Just this week I picked up an excellent hardback copy of same by internet auction for US $14.00. I also own the two earlier titles, and can attest that although there is a measure of duplication throughout the series each is sufficiently different to have on your book shelf. The last was a standard text at universities and colleges worldwide. It's out of print, remains in copyright (despite Terman's death three decades ago) but plenty of copies are available, see Amazon:

http://tinyurl.com/3e48po

I've already taken a bit of a punt in regard to copyright by uploading here. So here are a couple more chapters, (there is nothing like having a REAL copy to browse though!)

[Ears]

On negative feedback