Greetings Friends. I did a little amp swapping the other day and was surprised at the lack of bass coming from a Dynaco P-P amp I built a few years ago as compared to my daily amp, a 6B4G SET and was wondering why. Here's the schematic as built:
The amp in question was built using 2x PCBs with a circuit called "Dynaco 6V6 P-P" that features a 12AX7 in Floating Paraphase driving a pair of EL84 to about 10 watts. I used Edcor CXPP10-8k outputs with 4 and 8 ohm taps and wired the feedback to the 8 ohm tap. It was mentioned in a thread I started upon completion of the amp that the original Dynaco product used a different, higher-quality OT and adjustments to the Fb loop might be required. Using a simple scope and a 10kHz sq wave I minimized ringing by increasing the value of the cap in the Fb loop.
First listing to the amp gave a disappointing lack of bass when my Polk Monitor 60 speakers were connected to the 8 ohm taps, and better bass when wired to the 4 ohm taps - the Speakers show about 5 ohms of DC resistance from a DMM. But I get much more bass from my 6B4G SET amp hooked to those speakers - and that amp makes "only" 3 watts. So I'm wondering what I can do to improve the bass from this P-P amp. The original amp had the Fb loop connected to the 16 ohm tap, I only have 4 & 8 ohm - is this a problem? would changing the value of the Fb resistor compensate? Or do I just need new iron?
thanks for taking a look!
The amp in question was built using 2x PCBs with a circuit called "Dynaco 6V6 P-P" that features a 12AX7 in Floating Paraphase driving a pair of EL84 to about 10 watts. I used Edcor CXPP10-8k outputs with 4 and 8 ohm taps and wired the feedback to the 8 ohm tap. It was mentioned in a thread I started upon completion of the amp that the original Dynaco product used a different, higher-quality OT and adjustments to the Fb loop might be required. Using a simple scope and a 10kHz sq wave I minimized ringing by increasing the value of the cap in the Fb loop.
First listing to the amp gave a disappointing lack of bass when my Polk Monitor 60 speakers were connected to the 8 ohm taps, and better bass when wired to the 4 ohm taps - the Speakers show about 5 ohms of DC resistance from a DMM. But I get much more bass from my 6B4G SET amp hooked to those speakers - and that amp makes "only" 3 watts. So I'm wondering what I can do to improve the bass from this P-P amp. The original amp had the Fb loop connected to the 16 ohm tap, I only have 4 & 8 ohm - is this a problem? would changing the value of the Fb resistor compensate? Or do I just need new iron?
thanks for taking a look!
The best answer is that some measurments are needed to examine the frequency/power properties.
Those could be made on bothe the el84 and the 6bgg amps.
Those could be made on bothe the el84 and the 6bgg amps.
Maybe the 100uF capacitor in the input stage is the problem, try replacing it with more like 470uF at 16V or more.
Low voltage electrolytics don't last long, and the 100uF value is marginal at best.
Same goes for the capacitor across the output stage bias resistor, try 470uF at 50V there.
If your output tubes are not well matched for DC current at the bias point, that can also impair the bass response.
Try separate bias resistors of twice the resistance and at least 220uF capacitance each to have separate bias checks
for each output tube.
Low voltage electrolytics don't last long, and the 100uF value is marginal at best.
Same goes for the capacitor across the output stage bias resistor, try 470uF at 50V there.
If your output tubes are not well matched for DC current at the bias point, that can also impair the bass response.
Try separate bias resistors of twice the resistance and at least 220uF capacitance each to have separate bias checks
for each output tube.
Unfortunately the PCBs won't allow for separate K resistors, but the output tubes were matched by the seller.
I'll try 470R K resistors on v1b.
I'll try 470R K resistors on v1b.
You mean 470uF capacitor?
Instead of 6V, either 16V or 25V would be better there.
You'd be surprised by how little DC imbalance in the output tubes causes audible bass degradation.
Instead of 6V, either 16V or 25V would be better there.
You'd be surprised by how little DC imbalance in the output tubes causes audible bass degradation.
The input / pharaphase stage appears to be missing a cathode bias resistor which would be connected in parallel to the 100mFd, 6 volt capacitor.
Nothing is missing, that is a special topology used in the early days of Hafler.
http://www.r-type.org/articles/art-351.htm
http://www.r-type.org/articles/art-351.htm
Oh yea, caps in F. Somewhere I seem to recall reading that too high of a voltage rating on a bypass cap could lower the bass response, you are saying the opposite. Pretty sure I have some suitable caps, I'll try em tmrw.
@VaNarn the K resistor is all the way over on the Gnd side of the OT. That's what makes it Floating. It's biased thru the feedback loop.
The subjective impression of bass is not always explained in simple engineering terms, but some objective stuff, surprisingly works.
There is no excuse to not take some basic measurements. Anyone with a computer and a voltmeter can perform these.
- frequency response
- damping factor
An NFB free SET will easily have an order of magnitude worse Df than a PP amp with an overall NFB. In general the low damping factor will cause a substantial rise in bass response where the impedance peaks.
I have taken some measurements. SigGen set for 3vp-p, 1vrms.
Forgive the clumsy pics. I measured a sq wave at 1k, it yielded 3.3vp-p at the input.
After warming up the amp I applied the 1k 1vrms signal to input and got 8.86vrms measured at the output, across an 8 ohm resistor.
A 10kHz wave yielded 8.86vrms.
A 50Hz wave yielded 8.45vrms.
And a 20hz wave gave 7.25vrms.
So there's definitely a drop, I think I'll go back and see if I can plot the drop by measuring in 10hz intervals...
How do you measure damping factor?
Forgive the clumsy pics. I measured a sq wave at 1k, it yielded 3.3vp-p at the input.
After warming up the amp I applied the 1k 1vrms signal to input and got 8.86vrms measured at the output, across an 8 ohm resistor.
A 10kHz wave yielded 8.86vrms.
A 50Hz wave yielded 8.45vrms.
And a 20hz wave gave 7.25vrms.
So there's definitely a drop, I think I'll go back and see if I can plot the drop by measuring in 10hz intervals...
How do you measure damping factor?
That simply reflects the dominant low frequency pole, which is likely around 10Hz. Nothing unusual here.
If you increase the two electrolytic capacitors in the circuit each by a factor of five, (see post #4 above),
the pole will be lowered to around 2Hz, which is more standard.
If you increase the two electrolytic capacitors in the circuit each by a factor of five, (see post #4 above),
the pole will be lowered to around 2Hz, which is more standard.
The output Vp-p stays constant until I get below 60hz.
I have those caps on hand, I'll see what happens.
w
I have those caps on hand, I'll see what happens.
w
I apply a 100Hz sine wave and measure the output into two different loads, usually 4ohm and 8ohm. Then use a simple formula in a spreadsheet. The formula is derived from a basic equivalent circuit illustrating the amp output impedance in series with the load. There is a voltage divider between the amp's impedance and the load.
Using two loads is just one option, another is to take measurements with no load and a known load.
I somehow doubt the SET amp, provided it has no NFB to linearise the output transformer, will show a better LF response. Thanks to the gap most SE transformers are a little marginal with respect to inductance. Pure speculation of course, as we know nothing about the circuit or the transformer.