GU50 in a bass or guitar amp?

[imaradiostar]

If I'm not going for AB2 I've seriously considered using plate drive from a low Ra tube but I like the elimination of grid blocking the Ampeg design allows. Maybe I'll just use a driver transformer. I want to build a super simple but high output power section in the style of some of my guitar amps- there is something to the purity and simplicity that appeals to me, even if it's mostly in my head.

I've purchased mosfets for a similar design but again- you have to generate some specific supplies to make the best of that.

There is a part of me that wants to build a large monster amp with a built-in opto compressor like an LA2A. Ever since I saw studio bass rigs in Nashville with LA2A's I've thought that would be a neat idea. It would be pretty simple to add such a thing to a bass amp since the voltages and preamp are already there. I'd just need a little makeup gain (usually simple) and a 6aq5 or EL84 to drive the opto element.

jamie

[imaradiostar]

Good tip on the tek scope TF's by the way. I have transformers from Heathkit and some other giant kit scope. They're both around 15-16 pounds!

jamie

[thebasher]

Opto elements are in their nature non-linear which is probably why it is almost impossible to design FET ciruits for tremolo that sound like the old Fender amps. If you really want a good one look at Gyraf Audio. They have some interesting designs.
As for using driver/phase-splitter transformers: Not a bad idea, but expensive. I have attached Fender's 300PS schematic. You will run into terrible trouble if you apply a lot of negative feedback since phase shift through two transformer stages is unpredictable. You have been cautioned. Been there, done that and managed to jam the medium-wave band in a radius of 500 yards.
Basher
300ps.pdf

[imaradiostar]

That's pretty funny- I'd imagine you could generate RF like that for long before getting into trouble!

I'm not really a fan of negative feedback around the power amp so hopefully that won't be an issue. I'm looking for something that will transition from clean to dirty a little more gradually.

I guess with KT120's I could go ultralinear but not with the GU50 unless I have an output transformer with extra windings.

Gyraf stuff is pretty neat- in another life I worked in studios and I planned to build some of his designs but never got around to it. I have quite a few different LDR's to experiment with for an LA2A style optocoupler. As usual, too many projects and not enough time.

jamie

[thebasher]

There is a way to use a standard ultralinear tranny with tubes like GU50: Use a MOSFET to reduce the voltage and shunt the FET with a suitable MP capacitor in order to maintain the AC swing.

I use no-feedback amps for guitar because they distort in the perfect manner, but not for bass. The damping factor without NFB is too low and you rip your speaker's suspension due to wild uncontrolled exercions after transients.
If like me you bash your bass instead of picking it (hence my nickname) you need a good damping factor. Even small amps (especially small amps) can blow a speaker with 3-5 times the power rating. Not to speak of tweeters.

You have my heartfelt sympathy: So many projects, so little time. There are indeed certain limitations when it comes to tubeamps. The most significant is measured in WAL (Wife Acceptance Level). Wives do not see the beauty of a bright 845 triode lighting up the room and have no understanding for Klipsh horns and other decorative measures.

Basher

[imaradiostar]

That makes sense (mosfets for ultralinear). Would you vary the voltage around 300 or below 300 to not exceed the G2 voltage?

Good point on the NFB- I had assumed a stout enough amp wouldn't need NFB for bass but I'd imagine a bass cabinet pushes back pretty hard around resonance! This makes transformer coupling much more difficult. I suppose the SVT ended up the way it did for a reason.

My wife and I just had our first child. As you can imagine I haven't had much shop time lately!

jamie

[thebasher]

Congrats

Children sure take up much of your day. Enjoy them while they are small and you are still their hero.
In another amp design I actually measure MFB (motional feedback) from the speakers by introducing a 0.47 ohm resistor into the speaker circuit. The voltage drop across this resistor is fed to a balanced feedback circuit as an error signal compensating for the failing mechanical impedances of the speaker system.

I have attached a sample schematic for a PPP ultralinear amp delivering of app. 20W RMS The OPT is 6K primary but could be down to 4K for a few watts more - but who's counting. This amp was designed for HiFi using an open baffle speaker system and it sounds really good. It plays better with a simple two-way speaker system with the woofer connected directly and the tweeter via a simple crossover - preferably just a 2uF cap in series with the tweeter. Too complex crossovers may influence the MFB circuit - I think. The idea of using a current/voltage feedback is not mine - Philips used it in the late sixties and others have claimed ownership too. This design is field proven and works equally well with 4 x 6BQ5/EL84 or 4 x 6CA7/EL34 or 4 x 807.

The 20uH chokes are 100 ohm 2W resistors with about 20 turns of copper wire around them. Value is not critical as long as any oscillation is quenched. 20uH is a ballpark figure.
It is hard to calculate the damping factor in this circuit since all excessive cone movements are immediately compensated for. The damping is deteriorated by the resistor but apart from having a second voice coil on the woofer (as Philips did) I see no other way of measuring the counter motoric force from the speaker cone.
I have not yet tried this for a bass amp but I would imagine a significant improvement using 15" speakers with large cone mass.
Why 0.47 ohms in series? Dunno - seems to work fine but a bit of tweaking may even improve the circuit. However, the NFB in both sides should be roughly the same. The 200 ohms balance pot is for dynamic balance adjustment. It is very clear when you find the sweet point where the two circuit halves balance dynamically. For testing hook up a tone generator and run close to full power at 15Hz with an unmounted speaker as a load and you can actually see from the cone movement when the balance is optimal.

Basher


6V6PPP.PDF

[thebasher]

MOSFET in ultralinear

Yes, you are right: I set the g2 voltage to a steady 250VDC and use a 10uF MKH cap as an AC shunt. Just a gut feeling that setting the voltage to the max. at 300V is maybe cutting it a bit close with all the voltages farting around in an OPT.

[Steve Conner]

Great project, thebasher!

I've got one of those 300PS output transformers from Risson as well, but I was under the impression it was 3.3k to 8 ohms, with the secondaries in series, and the measurements I made support that. So are you just using it at half of its design impedance as a 1.65k to 4 ohm or do you have a different one?

I think the two transformers in the 300PS were purposely built with different low-end rolloff frequencies, to help loop stability at the LF end. The OT has a highish 50Hz spec, while the interstage is massively overbuilt with an enormous core. I don't suppose this helps the HF stability though, as you discovered.

So, running it as a 1.65k OT is probably a good thing, it should help extend the low end a little.

Anyway given that there's only a 4 ohm tap, how do you find the performance if you have to use the amp with an 8 ohm load? Do the GU50 screens light up?

[thebasher]

GU50's...

Hi Steve.

You are right about the transformer which is why I load thge secondary with 5.3 ohms. I am not pretending this is rocket science, just a ballpark effort and I report what I see/hear.
The ideal Raa for 6 x GU50 would be like 2.6k Ω in AB2, so putting a 5.3Ω load on a 8Ω winding is close enough.

The 1.65k I referred to was the Hammond transformer. It is a lot bigger than the Fender and has a better low-end. I have used it in the past for another project using 8 x EL34 and all you get from the extended lower range is subsonics – and who needs it. However, 1.65k Ω Raa will definitely produce the full 300W with Va at 560V.

You are probably right about the two transformers being purposely built like they are. Fender always ‘played it by ear’ (hence their impossible tone stack defying all logic) for the best subjective sound experience.

Example: Looking at the BluesDeVille guitar amp circuit + 4 pretty crummy 10” speakers there is not much to impress you. However, together with a 1960’s Telecaster – this is the sound as it should be. Or did we just grow up with it and got accustomed to it? Hard to tell.

On your suggestion I tried to run the amp with an 8 ohms load on the 4 ohms winding. Sounds OK, sound level somewhat lower as anticipated (depends on the speaker too) but no fireworks from the bottles. The damper diodes would take care of that – I hope. But still loud enough!

I would not recommend a lighter load, though. Too dangerous with potential flash-overs in the OPT. As you may have noticed the HT relay will not actuate if you pull the main speaker plug. A no-load condition at high volume would kill the tranny immediately, I am afraid.

When the transition to the new chassis is complete I shall publish some pictures of the nice glow from the GU50’s. You’ll love it. Unfortunately I cannot JPEG the sound…