After posting a question last month about the hammond 850Q, and not getting much response, I did some research on my own. I really was hoping to put together a transformer coupled amp, but it seems that there is almost nothing to be gained by doing so - and a lot to be potentially lost.
Am I wrong?
They seem to: a) cost more than many output transformers, b) have limited bandwidth - even the pricey ones, and c) have many restrictions as to input impedence, capacitance, plate current of the driver tube, plate resistance of the source, etc... Not to mention all the "specs" I see at most manufacturer's sites are for an unloaded condition. Is that a valid measurement?
None of these restrictions would apply if I were to simply rc couple it. Bandwidth can be as high as I want it to be.
Am I wrong?
They seem to: a) cost more than many output transformers, b) have limited bandwidth - even the pricey ones, and c) have many restrictions as to input impedence, capacitance, plate current of the driver tube, plate resistance of the source, etc... Not to mention all the "specs" I see at most manufacturer's sites are for an unloaded condition. Is that a valid measurement?
None of these restrictions would apply if I were to simply rc couple it. Bandwidth can be as high as I want it to be.
I have always wondered this myself- surely a capacitor, with its slight phase changes at different frequencies, would still have an order of magnitude less distortion than a non-linear device like a transformer?
One transformer in the circuit is probably enough! If you use a large enough coupling cap, the bandwidth would be very large.
Why then, apart from the simplicity factor, do people use them?
And why don't more people use direct coupling?
One transformer in the circuit is probably enough! If you use a large enough coupling cap, the bandwidth would be very large.
Why then, apart from the simplicity factor, do people use them?
And why don't more people use direct coupling?
Shifty, the usual reasons are fashion and desire to introduce euphonious (to the designer's ear) distortion and/or bandwidth limiting. Designs where the output is intended to accurately reflect the input do not generally use interstage transformers. Designs intended to "prettify" the sound often do.
Direct coupling in tube circuits can be quite successful, if thought through properly and set up with sufficient adjustments to reduce the impact of tube variations.
Direct coupling in tube circuits can be quite successful, if thought through properly and set up with sufficient adjustments to reduce the impact of tube variations.
Reason 1:
To drive a low mu triode, running at high voltage, the peak-to-peak drive required can exceed 400 volts. Try getting that out of a driver stage - and stay linear.
The transformer, even when 1:1 (the best bandwidth kind) allows twice the voltage swing.
NB An anode choke achieves the same thing in terms of drive voltage.
Reason 2:
To provide a linear load for the driver valve. This reduces distortion.
Reason 3:
To provide a low impedance AC source. This reduces distortion.
NB An anode choke in the driver achieves these things also.
These help satisfy the low natural distortion requirements for a no-feedback amplifier.
Indeed, global feedback would be most unwise with 2 transformers in the loop.
Cheers
To drive a low mu triode, running at high voltage, the peak-to-peak drive required can exceed 400 volts. Try getting that out of a driver stage - and stay linear.
The transformer, even when 1:1 (the best bandwidth kind) allows twice the voltage swing.
NB An anode choke achieves the same thing in terms of drive voltage.
Reason 2:
To provide a linear load for the driver valve. This reduces distortion.
Reason 3:
To provide a low impedance AC source. This reduces distortion.
NB An anode choke in the driver achieves these things also.
These help satisfy the low natural distortion requirements for a no-feedback amplifier.
Indeed, global feedback would be most unwise with 2 transformers in the loop.
Cheers
John, I don't want to see the triode that needs a grid voltage of minus 200volts!!! Yikes! 
As far as your point that the IT provides lower distortion - the few times distortion measurements were given for IT's, it's always around 1.5% - which is about what a typical RC stage is, before NFB. Isn't it?
The higher swing argument is easy to understand.
As far as your point that the IT provides lower distortion - the few times distortion measurements were given for IT's, it's always around 1.5% - which is about what a typical RC stage is, before NFB. Isn't it?
The higher swing argument is easy to understand.
Joel,
The distortion argument is that the transformer allows the preceeding and following stage to run at lower distortion.
I don't have any figures for the actual distortion within the transformer; where did you find those figures?
IMO a capacitor and 2 resistors on their own create practically no distortion. But I'd better keep quiet about that
Cheers,
The distortion argument is that the transformer allows the preceeding and following stage to run at lower distortion.
I don't have any figures for the actual distortion within the transformer; where did you find those figures?
IMO a capacitor and 2 resistors on their own create practically no distortion. But I'd better keep quiet about that
Cheers,
I'll add Reason #4:
An interstage tranny secondary provides a low DCR path to ground as compared to a grid resistor on the driven (ie, output) tube. Less concerns with grid current in that tube.
In addition, the secondary of the interstage can be a fine place to inject the voltage required for fixed bias of the output tube, if your tastes run to that method. - Pat
An interstage tranny secondary provides a low DCR path to ground as compared to a grid resistor on the driven (ie, output) tube. Less concerns with grid current in that tube.
In addition, the secondary of the interstage can be a fine place to inject the voltage required for fixed bias of the output tube, if your tastes run to that method. - Pat
GOLDEN WATT
Hi,
That's one reason you often only find IS Xformers in low wattage designs.
Other than that I think Joels' got a valid point even though biasing tubes in the very negative voltage range has nothing frightening to me.
Just pick the tube for the job and you're done.
Ciao,
P.S.Corbieres tonight.
Hi,
That's one reason you often only find IS Xformers in low wattage designs.
Other than that I think Joels' got a valid point even though biasing tubes in the very negative voltage range has nothing frightening to me.
Just pick the tube for the job and you're done.
Ciao,
P.S.Corbieres tonight.
Frank,
As you might have noticed from my IP, I was at work. So I think the earlier Corbieres taunt was below the belt!
I'm home now, but dry tonight.
Of course IS transformers or CR are "horses for courses".
It would be counter-productive using an IS in the EL34 amps now under disussion in the other thread.
Good-night,
As you might have noticed from my IP, I was at work. So I think the earlier Corbieres taunt was below the belt!
I'm home now, but dry tonight.
Of course IS transformers or CR are "horses for courses".
It would be counter-productive using an IS in the EL34 amps now under disussion in the other thread.
Good-night,
tubesguy said:
Yeah, I thought that would be nice. I want to use fixed bias on the UX-201A's.
Anyway, if I just RC couple two of them, with 22k plate resistors, I can get a gain of 36, which is getting near a reasonable input sensitivity for the amp.
Or, maybe I should just give up on the antiques.
I don't have any one source Frank, but since I'm often willing to risk it and buy "untested" tubes, I rarely pay more than $5 US for anything. So, now I have a shoebox full of globe 01-A's sitting next to the one full of 6SN7's... God, I love eBay.
John - they will be the two gain stages in a PSE 71-a monoblock (maybe). Volume control after the first 01A. Filaments will be DC, using the rectifier circuit from my 6F5/71A circuit I posted before. It works really well. I also like not having to mess with balance pots and all that. Makes for a clean schematic.
I breadboarded a single stage last year, and was happy with the performance - even if the volume was too low. I think cascaded stages will solve that issue.
Joel
dhaen said:
John - they will be the two gain stages in a PSE 71-a monoblock (maybe). Volume control after the first 01A. Filaments will be DC, using the rectifier circuit from my 6F5/71A circuit I posted before. It works really well. I also like not having to mess with balance pots and all that. Makes for a clean schematic.
I breadboarded a single stage last year, and was happy with the performance - even if the volume was too low. I think cascaded stages will solve that issue.
Joel
IS Transformer distortion
Joel,
A transformer is on the bench in front of me now.
I've got nearly 15V RMS on it's input. That is just below the IS voltage required in my amp to get 1 watt output, but the most I can get without having to introduce "unknowns".
The THD measures 0.02%. That's just above the lowest I can read (0.015); due to equipment limitations.
Since the transformer is "gapped" for 40mA, for use in SE, I'd better introduce some DC:
With 15mA DC, which is the most I can easilly get from a 9v battery without oscillator loading:
The THD measures 0.025%. There are extra lengths of wire, and a DVM in circuit now, so I would give there figures some tollerance.
Result:
At the power I'm most interested, there is no significant measurable distortion introduced by my IS transformer.
My ears had told me this already, but it is satisfying when the 2 test methods agree.
No doubt the large signal figures will be considerably "worse".
Until I finish my chassis, I can't properly measure the overall low power distortion.
Now can I get back to my chassis metalwork PLEASE
Cheers,
Joel,
A transformer is on the bench in front of me now.
I've got nearly 15V RMS on it's input. That is just below the IS voltage required in my amp to get 1 watt output, but the most I can get without having to introduce "unknowns".
The THD measures 0.02%. That's just above the lowest I can read (0.015); due to equipment limitations.
Since the transformer is "gapped" for 40mA, for use in SE, I'd better introduce some DC:
With 15mA DC, which is the most I can easilly get from a 9v battery without oscillator loading:
The THD measures 0.025%. There are extra lengths of wire, and a DVM in circuit now, so I would give there figures some tollerance.
Result:
At the power I'm most interested, there is no significant measurable distortion introduced by my IS transformer.
My ears had told me this already, but it is satisfying when the 2 test methods agree.
No doubt the large signal figures will be considerably "worse".
Until I finish my chassis, I can't properly measure the overall low power distortion.
Now can I get back to my chassis metalwork PLEASE
Cheers,
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.