crt said:
Crt, if symasym oscillates, it sounds bad. It will be lifeless, no dynamics, no bass and muffled.
Sadly, you have omitted R16/19, with these you would have been absolutely sure about stability.
May i suggest, but only if you use at most +/-36v supply, that you use 2sk170 as input devices? These have half the gain (lowering the feedback, increasing stability) and sound better anyway.
Dexter, you have definitely something wrong with your vbe-multiplier. Maybe you got the pinout of the pot wrong, having a constant resistance here? Turning the pot must show ANY difference.
Mike
Andrew, Chris, i also have the experience that hfe matched devices automatically match vbe. When i select devices, i always measure vbe and hfe. With the onsemi devices, i have a good match after measuring ~5 devices. (< 1%)
For the outpudevices vbe match is not necessary, the few mv dc-offset introduced by vbe mismatch here is easily compensated by feedback.
Mike
For the outpudevices vbe match is not necessary, the few mv dc-offset introduced by vbe mismatch here is easily compensated by feedback.
Mike
Hi Mike,
I think we all agree. Andrew may have some mixed batches there. In that case, he would have to be careful of both. So he is correct in his situation.
Hey Mike! I just got an HP 3585A for my bench. 😀 😀 😀 😀
Hi Andrew,
I will sometimes separate my transistors into rough beta groups, then simply match pairs using the jig. That's all you should need to do. These parts are cheap enough that unmatched parts get used by themselves. You can always buy a 100 pc lot for $20 ~ $30. Then you are working with the same lot (hopefully). Depends on the vendor. You should be able to request the same lot. If they can't supply that way - change vendors.
-Chris
I think we all agree. Andrew may have some mixed batches there. In that case, he would have to be careful of both. So he is correct in his situation.
Hey Mike! I just got an HP 3585A for my bench. 😀 😀 😀 😀
Hi Andrew,
Actually, you often do not need to know the exact value of beta. If you measure across the collectors (from one to the other transistor) and get a few mV DC, that's a nicely matched pair.
I will sometimes separate my transistors into rough beta groups, then simply match pairs using the jig. That's all you should need to do. These parts are cheap enough that unmatched parts get used by themselves. You can always buy a 100 pc lot for $20 ~ $30. Then you are working with the same lot (hopefully). Depends on the vendor. You should be able to request the same lot. If they can't supply that way - change vendors.
-Chris
No it isn't.anatech said:
A different beta (hFE) will draw a different base current when the Ic values are precisely matched. The different base currents will force the two transistors into different Vbe since their emitters are connected.
I am now beginning to wonder if your test procedure has overlooked what the test circuit is allowing you to measure.
I got my bc546c and bc556c from the 10,000 lot that was organised as a group buy. Did we have a choice of vendor?anatech said:
Hi Andrew,
The two bases are connected to common through matched resistors. The emitters are connected together to a current source. Each collector connects to the Vcc supply through matched resistors.
So, if your collector voltages are equal (voltage drops are equal), then your currents must be equal. The emitter voltages are locked together. The B-E drops will be very, very close as long as the devices are at the same temperature and from the same lot. That's just how things are. Of course, if you hang a meter between the bases, you should measure something in the hundred uV range I would think. Easily checked. In this way you could match between lots as long as you checked at different currents.
Is this how you understand the jig to work Andrew?
-Chris
No idea. I tend to buy in 100 lots.
The way mine works is simple. It mimics an actual circuit.
The two bases are connected to common through matched resistors. The emitters are connected together to a current source. Each collector connects to the Vcc supply through matched resistors.
So, if your collector voltages are equal (voltage drops are equal), then your currents must be equal. The emitter voltages are locked together. The B-E drops will be very, very close as long as the devices are at the same temperature and from the same lot. That's just how things are. Of course, if you hang a meter between the bases, you should measure something in the hundred uV range I would think. Easily checked. In this way you could match between lots as long as you checked at different currents.
Is this how you understand the jig to work Andrew?
-Chris
Hi Anatech,
yes, that's how I understand the jig you posted works and how I built it.
I soon realised that when one measures the voltage across the base resistors they were different and this difference can be measured directly across the base pins.
It is this base voltage difference that indicates the different hFE and when hFE is different the applied Vbe MUST BE DIFFERENT.
That is the fundamental difference in our approach. I measure Vbe and/or the base voltage difference and set this as the more important parameter in obtaining the matched Ic, again by measuring the collector voltage difference.
It is to avoid the base voltage difference that I then shorted out the base resistors once I had identified fairly close matches. If Ic with base resistors in circuit matches and Ic still matches when the base resistors are shorted out then the two test transistors match for both Vbe and hFE.
yes, that's how I understand the jig you posted works and how I built it.
I soon realised that when one measures the voltage across the base resistors they were different and this difference can be measured directly across the base pins.
It is this base voltage difference that indicates the different hFE and when hFE is different the applied Vbe MUST BE DIFFERENT.
That is the fundamental difference in our approach. I measure Vbe and/or the base voltage difference and set this as the more important parameter in obtaining the matched Ic, again by measuring the collector voltage difference.
It is to avoid the base voltage difference that I then shorted out the base resistors once I had identified fairly close matches. If Ic with base resistors in circuit matches and Ic still matches when the base resistors are shorted out then the two test transistors match for both Vbe and hFE.
Mike,
Greetings.
Good and bad news.
after setting aside my symasym for a couple of days
today, I opened the case and measured again .....
now, I'm able to set and measure 2.4 v between driver base U5 and U6,
However the measured bias won't stabilize, the value goes up and down errratically.
I checked and tightened T1 but to no avail.
Please advice what to do?
many thanks and best rgds,
Greetings.
Good and bad news.
after setting aside my symasym for a couple of days
today, I opened the case and measured again .....
now, I'm able to set and measure 2.4 v between driver base U5 and U6,
However the measured bias won't stabilize, the value goes up and down errratically.
I checked and tightened T1 but to no avail.
Please advice what to do?
many thanks and best rgds,
Hi Andrew,
Okay, we are basically saying the same thing. You can measure your E-B voltage drops to the same effect. Shorting out those resistors really only makes it impossible to read the E-B voltages, but it does force them to be equal. This will force different base currents (greater magnitude) and amplify the differences. I simply check the values of E-B voltages.
Just thinking here, I normally use Japanese signal transistors. They are generally more consistent than the US types and probably the Pro Electron type. E-B variations may possibly be greater for you. I am guessing at this simply because I don't see the problem that you are describing. Maybe I should match some JAN devices and see.
Hi dexter,
Check solder joints in the bias area. You may have a cracked bias transistor there. They only need to be snug, not tight.
-Chris
Okay, we are basically saying the same thing. You can measure your E-B voltage drops to the same effect. Shorting out those resistors really only makes it impossible to read the E-B voltages, but it does force them to be equal. This will force different base currents (greater magnitude) and amplify the differences. I simply check the values of E-B voltages.
Just thinking here, I normally use Japanese signal transistors. They are generally more consistent than the US types and probably the Pro Electron type. E-B variations may possibly be greater for you. I am guessing at this simply because I don't see the problem that you are describing. Maybe I should match some JAN devices and see.
Hi dexter,
Check solder joints in the bias area. You may have a cracked bias transistor there. They only need to be snug, not tight.
-Chris
Dexter, how much does the bias vary? Can you show a photo of your construction? There might be a layout issue.
Of course, it is not normal that a circuit becomes functional on its own. You might consider replacing T1.
What are you using as output devices? Are the drivers on the same heat sink as the output devices?
An erratically changing bias can also be caused by oscillation. Do you have a scope at hand?
Mike
Of course, it is not normal that a circuit becomes functional on its own. You might consider replacing T1.
What are you using as output devices? Are the drivers on the same heat sink as the output devices?
An erratically changing bias can also be caused by oscillation. Do you have a scope at hand?
Mike
Thanks Mike and Anatech for your reply.
Please find my symasym layout.
Please do look at it.
many thanks and best rgds,
Please find my symasym layout.
Please do look at it.
many thanks and best rgds,
Mike ,
The bias vary between 2-100+ mv
I'm using 2sc/a 5200/1943
driver 1837/4793
t1: bd139
all to92 tr are the same type except mpsa18 -> BC 550C
22-0-22V ac transformer
best rgds,
The bias vary between 2-100+ mv
I'm using 2sc/a 5200/1943
driver 1837/4793
t1: bd139
all to92 tr are the same type except mpsa18 -> BC 550C
22-0-22V ac transformer
best rgds,
dexter said:
😕 have you missed pnp/npn ?
bc550c has max rating of 45v, WAY TOO LOW for symasym!!!
2n5401/2n5551 is not subject to be substituted.
Mike
Sorry Mike,
What I mean was all TO92 transistor are the same type as per yr BOM, except MPS A18 -> BC 550c.
I think I found the problem, I change the trimmer + C5 and
my symasym Vbe is working now
the bias's set at 55mV and rock stable after 5 minutes warm up.
the amplifier's been on for more than 6 hrs now
Dc offset is 0.5 mV and 1.2 mV
however (i'm measuring with my hand)
the heatsink temperature is different between
two channel, one is hotter than the other. bias was measurd the same for both channel.
my impression :
the amplifier sounds pleasant , almost no harshness and etched details clean midrange and slightly soft treble.
bass is not as pronounced and punchy as my other low power power amp (esp3a, Gain clone, a couple of mosfet amp)
my symasym psu caps 6,800uf X4 for each channel
soundscape and separation between musical instrument ''s quite good.
Mike,
I hope that I didn't annoy you with another questions
can I use 2sk30 for i/p?
How to reduce symasym's gain?
i'm using tube preamplifier and the hiss's clearly audible,
Which it's not so with other power amplifier.
Many thanks Mike for your kind help,
best rgds,
What I mean was all TO92 transistor are the same type as per yr BOM, except MPS A18 -> BC 550c.
I think I found the problem, I change the trimmer + C5 and
my symasym Vbe is working now
the bias's set at 55mV and rock stable after 5 minutes warm up.
the amplifier's been on for more than 6 hrs now
Dc offset is 0.5 mV and 1.2 mV
however (i'm measuring with my hand)
the heatsink temperature is different between
two channel, one is hotter than the other. bias was measurd the same for both channel.
my impression :
the amplifier sounds pleasant , almost no harshness and etched details clean midrange and slightly soft treble.
bass is not as pronounced and punchy as my other low power power amp (esp3a, Gain clone, a couple of mosfet amp)
my symasym psu caps 6,800uf X4 for each channel
soundscape and separation between musical instrument ''s quite good.
Mike,
I hope that I didn't annoy you with another questions
can I use 2sk30 for i/p?
How to reduce symasym's gain?
i'm using tube preamplifier and the hiss's clearly audible,
Which it's not so with other power amplifier.
Many thanks Mike for your kind help,
best rgds,
Hi Dexter, sorry for completely misunderstanding you...
Nice to hear that you fixed the problem.
For lowering the gain you need to increase r30 from 500ohm to 1k, and add emitter resistors to q1/q2 (-> r16/19). Without these emitter resistors symasym would start to oscillate.
SK30 should work as input device, however, i recommend sk170 as input device for jfet.
Is the hiss coming from the preamp or from symasym?
Mike
Nice to hear that you fixed the problem.
For lowering the gain you need to increase r30 from 500ohm to 1k, and add emitter resistors to q1/q2 (-> r16/19). Without these emitter resistors symasym would start to oscillate.
SK30 should work as input device, however, i recommend sk170 as input device for jfet.
Is the hiss coming from the preamp or from symasym?
Mike
I use BC550 for input, and I use +/-32V at the supply.
All the parts as same as in your homepage.
I use soundcard (computer) for the sources, should I use R16/R19 and raise the 500R to 1K?
All the parts as same as in your homepage.
I use soundcard (computer) for the sources, should I use R16/R19 and raise the 500R to 1K?
No, don't change 500ohm to 1k, without preamp symasym's gain fits quite good. The sk170 was just a suggestion, and is a direct drop in.
Mike
Mike
Mike,
Hreetings.
I haven't got achance to try the fet and lowering the fgain.
Will inform you when i've done so.
The hiss was coming from the preamp,
unconnected to preamp
my symasym is silent except for very slight hum.
Hreetings.
I haven't got achance to try the fet and lowering the fgain.
Will inform you when i've done so.
The hiss was coming from the preamp,
unconnected to preamp
my symasym is silent except for very slight hum.
HI Mike B
can i use bigger/lower value for pot. (maybe 10K,25K, 50, or 100K) for volume control? Very hard to find 22K pot. what your suggestion value.
until now i connect my symasym direct to soundcard without pot.
can i use bigger/lower value for pot. (maybe 10K,25K, 50, or 100K) for volume control? Very hard to find 22K pot. what your suggestion value.
until now i connect my symasym direct to soundcard without pot.