My little SE el84 DC amp. in my workshop gave me a suprise recently! Because the mains transformer is not well regulated, and to avoid stressing the smoothing and reservoir psu caps., I fitted a switch in the ht line to provide surge limiting, with a 10k ww resistor in circuit at switch on. I powered- up the amp. and was casually listening to a quiet musical passage, and remember thinking how 'good' it sounded, infact I took a more concentrated listen for a few moments which confirmed the first impression. Only when a crescendo arrived did I notice some distortion! I checked and yes, I had forgotten to throw the standby switch - which I then did. I started the cd again but was aware that it appeared not to sound as good as at first listening. I wondered if feeding a reservoir cap. via a high resistance (10k) would effectively give the better results of my first listening. I therefore fitted a large value reservoir (2200uf 330v) and was pleased to confirm the apparent increase in fidelity. I appreciate that using a high value feed resistor ruins the psu's impedance. I wonder if anyone else uses this technique of 'trickle charging' a large reservoir cap? My amp has modest current requirements at full power - maybe 90mA or so which appears to be within the storage capacity of the capacitor, sufficient at least to cope with crescendo,s without suffering a voltage drop. Any comments welcome please.
regards,
David.
regards,
David.
An interesting observation which will no doubt annoy the 'engineering' types 🙂
The PS impedance issue is far from obvious. The classic approach advocates large PS caps and local decoupling, thus rendering the transformer side of the impedance equation unimportant.
Another approach based on the presumption that large capacitors are evil requires low impedance transformers, low impedance chokes and low impedance rectifiers in order to get away with tiny caps.
I would normally prefer a low cap design although once, out of necessity, i tried something similar to your ps. It was a phono pre with a tube rectifier and a huge 3300uF main cap being trickle charged in order to keep the rectifier alive. Contrary to expectations it didn't sound dead or undynamic although the bass was certainly not very fast.
It may well work with a small power amp but you obviously need to ensure that under max load the caps manage to recharge sufficiently. PS designer will give you a quick answer.
The PS impedance issue is far from obvious. The classic approach advocates large PS caps and local decoupling, thus rendering the transformer side of the impedance equation unimportant.
Another approach based on the presumption that large capacitors are evil requires low impedance transformers, low impedance chokes and low impedance rectifiers in order to get away with tiny caps.
I would normally prefer a low cap design although once, out of necessity, i tried something similar to your ps. It was a phono pre with a tube rectifier and a huge 3300uF main cap being trickle charged in order to keep the rectifier alive. Contrary to expectations it didn't sound dead or undynamic although the bass was certainly not very fast.
It may well work with a small power amp but you obviously need to ensure that under max load the caps manage to recharge sufficiently. PS designer will give you a quick answer.
Single ended amplifier operates class a, current demand should not change significantly at any signal level at which the amplifier operates linearly. The comment you made probably indicates that the crescendo clipped the amplifier at a level where that would not normally happen. I think what might be happening is that your plate voltage is much lower in the "standby" position and you have found a new, perhaps more musical operating point as a consequence of the lower voltage.
I would measure the plate voltage and the cathode current - that should tell you a lot. It may be that you can choose another operating point that sounds preferable to you.
I spent a long time experimenting with the operating point of the output stage in my very first 45 based SE amplifier. The conditions that produced the most power did not to me sound the best.
Kevin
www.kta-hifi.net
I would measure the plate voltage and the cathode current - that should tell you a lot. It may be that you can choose another operating point that sounds preferable to you.
I spent a long time experimenting with the operating point of the output stage in my very first 45 based SE amplifier. The conditions that produced the most power did not to me sound the best.
Kevin
www.kta-hifi.net
Something similar.......
/*Just completed a bread-board OTL. of the circlotron type
The floating supplies are coming from a pair of voltage-doublers using 3300uF caps and a couple of diodes with a toroid 55V,55V @6A. as this transformer is all I had around...
For safety, as my diodes are a little small on current rating for this application, and the caps had been stored for around 10years, I placed a couple of 2.7 ohm WW resistors in series with the 55V transformer leads to reduce inrush currents, One in each secondary... My floating supplies off-load derive 160V. On load at 150mA bias, we had 155V
During sound testing, I was fairly impressed, untill louder music was played, so I did some tests, where I found that above around 20W I was getting 'crossover-distortion'
Of course what was happening, under load the floating supplies were dropping considerably as they were called to supply over 2A and relieving the bias from the O/P pair, and hence the Xover distortion at higher o/p levels.--That confused me for a while untill I checked the supply rails to the o/p pair........
Shorting out the 2.7ohm resistors transformed the sound and power, as well as an additional 3300uF cap on each supply rail!
.
/*Just completed a bread-board OTL. of the circlotron type
The floating supplies are coming from a pair of voltage-doublers using 3300uF caps and a couple of diodes with a toroid 55V,55V @6A. as this transformer is all I had around...
For safety, as my diodes are a little small on current rating for this application, and the caps had been stored for around 10years, I placed a couple of 2.7 ohm WW resistors in series with the 55V transformer leads to reduce inrush currents, One in each secondary... My floating supplies off-load derive 160V. On load at 150mA bias, we had 155V
During sound testing, I was fairly impressed, untill louder music was played, so I did some tests, where I found that above around 20W I was getting 'crossover-distortion'
Of course what was happening, under load the floating supplies were dropping considerably as they were called to supply over 2A and relieving the bias from the O/P pair, and hence the Xover distortion at higher o/p levels.--That confused me for a while untill I checked the supply rails to the o/p pair........
Shorting out the 2.7ohm resistors transformed the sound and power, as well as an additional 3300uF cap on each supply rail!
.
Thanks to all who replied to this thread. My OTL also will work at low level with much reduced ht - but can't handle a crescendo!
Thanks again,
Regards,
David.
Thanks again,
Regards,
David.
OTL
I was getting around 15V p-p O/P accross an 8 ohm load before I got that odd Xover distortion, so something like 30W P-P, tops
After shorting the 2.7ohm resistors, my output was up to 28V p-p accross the same 8 ohm. something like 90 odd Watts, P-P A three fold improvement!, where it just goes into clipping at the peaks of the sine, more or less simmetrically.
Not bad for only 2 unmatched, El Cheapo 6C33C-B, Roughly 3.5A through them!--
Sounds quite amazing now, Very powerful bass, but has the great clarity the OTL is famous for!
When I get some time to do some more testing, Ill find how much the HT rails were actually dropping to with those resistors in circuit!
I was getting around 15V p-p O/P accross an 8 ohm load before I got that odd Xover distortion, so something like 30W P-P, tops
After shorting the 2.7ohm resistors, my output was up to 28V p-p accross the same 8 ohm. something like 90 odd Watts, P-P A three fold improvement!, where it just goes into clipping at the peaks of the sine, more or less simmetrically.
Not bad for only 2 unmatched, El Cheapo 6C33C-B, Roughly 3.5A through them!--
Sounds quite amazing now, Very powerful bass, but has the great clarity the OTL is famous for!
When I get some time to do some more testing, Ill find how much the HT rails were actually dropping to with those resistors in circuit!
Hi Alastair,
I assume you meant 28Vpk? Power is either specified in terms of rms or pk, not peak to peak. If indeed it were pk-pk then the actual output power before clipping is 14Vp sqd/8 ohms which is slightly under 25Wpk and or equivalent to about 12.0Wrms.
Kevin
I assume you meant 28Vpk? Power is either specified in terms of rms or pk, not peak to peak. If indeed it were pk-pk then the actual output power before clipping is 14Vp sqd/8 ohms which is slightly under 25Wpk and or equivalent to about 12.0Wrms.
Kevin
- Status
- Not open for further replies.