I have really tried to sort this out without asking for advice, but I don’t know where to turn any more. I am doing everything I can think of to linearize/stabilize the mosfets in my hybrid amp. So far I have done the following:-1) Fiddled with gate stopper values. 2) Put caps right on the drain pins where the power supply is connected. 3) Output inductor 4) Zobel There has been a lot of improvement from these mods, but there is still a slight distortion apparent in certain passages with complex waveforms (FWIW, the biggest improvement I noticed was due to the PS bypass caps).
Can anyone suggest anything else I can try?
Can anyone suggest anything else I can try?
The distortion is like a kind of haze/buzz around piano notes. It varies in intensity according to the frequency of the note. It is not a loud/obvious distortion. You have to put your ear near the speaker to hear it. It is also not noticeable most of the time, but there is a particular passage of piano with a lot of reverb that I am using as a reference to test my mods. (It isn't the recording. My Hart Linsley Hood amp which uses the same mosfets doesn't produce the haze in this passage)
As diy4 suggests, need to see the circuit.
I've also seen oscillation cause haze/fuss on certain types of sound. Piano notes have fast edges and this can set it off, especially into capacitive loads. Don't put a cap across the output as a test unless you've got some way to protect the output stage from melt-down.
If it is instability then there are many potential causes, some of which have nothing to do with the schematic. Ground paths are a notorious cause.
I've also seen oscillation cause haze/fuss on certain types of sound. Piano notes have fast edges and this can set it off, especially into capacitive loads. Don't put a cap across the output as a test unless you've got some way to protect the output stage from melt-down.
If it is instability then there are many potential causes, some of which have nothing to do with the schematic. Ground paths are a notorious cause.
I will see what i can do for a drawing with voltages. In the meantime I can tell you it is a single supply rail, so has DC blocking output caps. (Phillips 4700uF bypassed with 0.47 polyprops) I have put a zobel, 10r and .15uF polyprop, after the caps and before the inductor. I may remove the zobel but i am just giving it a chance to settle first before making a decision on that.
I am also thinking that it may be some sort of ringing that i am hearing which is being triggered by complex waveforms. If it is crossover distortion, then why is it so seldom audible?
I am also thinking that it may be some sort of ringing that i am hearing which is being triggered by complex waveforms. If it is crossover distortion, then why is it so seldom audible?
Hi Andrew. The output stage puts out way too much power to be class A (except for the first watt or two). I also recall the guy who sold it to me showing me the frequency response graph with the typical class B "wings" which he was at pains to explain. I think this amp is a combination of ideas. I would very much like to know where the designs originated from that led to this amp. I am pretty sure they would have been seen in magazines about 10 to 12 years ago. Any ideas?
Some measurements. B+ is 94v. Input is SSM2210 (1 half per channel). Emitter = 0v, base .6v, collector 3.2v. Gain stage EL34, triode strapped, Output, a pair of mosfets. Gates sit at around 42v,
as do the anodes of the el 34's. Drain of n-channel at full B+. I can't quite figure out the arrangement for feeding the correct power to the various devices. Does this make sense to anyone?
Some measurements. B+ is 94v. Input is SSM2210 (1 half per channel). Emitter = 0v, base .6v, collector 3.2v. Gain stage EL34, triode strapped, Output, a pair of mosfets. Gates sit at around 42v,
as do the anodes of the el 34's. Drain of n-channel at full B+. I can't quite figure out the arrangement for feeding the correct power to the various devices. Does this make sense to anyone?
Hi Andrew,
no, the fets are not fed from a transformer. From memory, gate voltages are all the same (between 41 - 44v). At the moment, the amp is not working. (see post a few back). The power seems normal to all devices, so i guess I must have damaged the mosfets. I had a cap on the gate of an n mosfet and shorted it while trying to solder a 150pf cap from gate to drain. This must have unloaded whatever charge the cap was still carrying (anything up to about 90v) into the earth circuit. So i guess this would have applied a serious reverse voltage into the drains of the P channel mosfets. Do you think it would have damaged the n-channel devices as well? (I am having serious dificulty trying to ascertain if mosfets are damaged with only a cheap DMM at my disposal.)
no, the fets are not fed from a transformer. From memory, gate voltages are all the same (between 41 - 44v). At the moment, the amp is not working. (see post a few back). The power seems normal to all devices, so i guess I must have damaged the mosfets. I had a cap on the gate of an n mosfet and shorted it while trying to solder a 150pf cap from gate to drain. This must have unloaded whatever charge the cap was still carrying (anything up to about 90v) into the earth circuit. So i guess this would have applied a serious reverse voltage into the drains of the P channel mosfets. Do you think it would have damaged the n-channel devices as well? (I am having serious dificulty trying to ascertain if mosfets are damaged with only a cheap DMM at my disposal.)
Hi hihopes
I think your amp is based on a Jeff Macaulay design published in Practical Electronics September 1991. Jeff has a website at www.inline-electronics.co.uk. He appears to be offering an updated version for sale as a kit.
HTH.
Steve.
I think your amp is based on a Jeff Macaulay design published in Practical Electronics September 1991. Jeff has a website at www.inline-electronics.co.uk. He appears to be offering an updated version for sale as a kit.
HTH.
Steve.
As I understand it, it is a class A design, but the biasing current is not high enough to keep it in class A for more than the first watt or two. The gate of the n-channel is fed from the anode of the valve. The other gate is fed with power from a regulation circuit of some sort. Does this make sense to you? The voltages are not exactly the same on the n and p gates. Should they be?
Hi hihopes,
The schematic for your amp is not available on his site. It is an 80w class AB. The bias is set by a 150R resistor between the gates of the fets. The original design used 2sk135 and 2sj50 fets. The gate of the p-channel fet is connected to the junction of the el34 anode and the 150R bias resistor, the gate of the n-channel fet is connected to the junction of the 150R bias resistor and an 8k2 anode load resistor fed with approx 95v from a two transistor regulator. There should be a preset pot to set this voltage.
HTH.
Steve.
The schematic for your amp is not available on his site. It is an 80w class AB. The bias is set by a 150R resistor between the gates of the fets. The original design used 2sk135 and 2sj50 fets. The gate of the p-channel fet is connected to the junction of the el34 anode and the 150R bias resistor, the gate of the n-channel fet is connected to the junction of the 150R bias resistor and an 8k2 anode load resistor fed with approx 95v from a two transistor regulator. There should be a preset pot to set this voltage.
HTH.
Steve.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.