Curious, which models or series are you referring to? As Counterpoint products were high-end I'm surprised they didn't put source resistors in a commercial product, even I as a hobbyist quickly realized how essential those are...
I was warranty for Counterpoint. The SA series mosfet amps. They didn't do well with the later BJT amps either.
With Counterpoint, their design prowess with solid state was pretty poor and driven mostly by fads, as were his later "upgrades". I redesigned almost all the preamp power supplies to be reliable and less noisy (plus more stable in voltage). Other "Counterpoint experts" did terrible things to these units, and did not improve reliability. I get them from the USA also. It isn't uncommon for one to be repaired 3 ~ 4 times before I see it, and by then their owners have either given up or are desperate.
You know, the term "high end" was mostly self designated by the manufacturer itself. Several other brands considered "mid-Fi" performed a great deal better and were definitely more reliable. A redesigned Counterpoint preamp is pretty quiet with very low distortion. You would want an SA-3000 or 5000 over a 3.1 or 5.1 - easily.
With Counterpoint, their design prowess with solid state was pretty poor and driven mostly by fads, as were his later "upgrades". I redesigned almost all the preamp power supplies to be reliable and less noisy (plus more stable in voltage). Other "Counterpoint experts" did terrible things to these units, and did not improve reliability. I get them from the USA also. It isn't uncommon for one to be repaired 3 ~ 4 times before I see it, and by then their owners have either given up or are desperate.
You know, the term "high end" was mostly self designated by the manufacturer itself. Several other brands considered "mid-Fi" performed a great deal better and were definitely more reliable. A redesigned Counterpoint preamp is pretty quiet with very low distortion. You would want an SA-3000 or 5000 over a 3.1 or 5.1 - easily.
Dang... According to a Stereophile article an SA-100 would retail 1295USD in 1991 which would be close to 3 grand in today's USD... The Stereophile review was devastating, one of the most negative reviews I ever read in a magazine...
"I absolutely could not believe how flawed the SA-100 sounded, from top to bottom"
Stereophile SA-100 review
I already knew that brand names don't guarantee good sound and one has got to do their homework and read reviews, but those are still subjective and nowadays it's hard to find brick and mortar hi-fi stores where you can actually listen to products. Feels like the customer is left to buy half-blind...
"I absolutely could not believe how flawed the SA-100 sounded, from top to bottom"
Stereophile SA-100 review
I already knew that brand names don't guarantee good sound and one has got to do their homework and read reviews, but those are still subjective and nowadays it's hard to find brick and mortar hi-fi stores where you can actually listen to products. Feels like the customer is left to buy half-blind...
Stereophile was bang-on.
I redesigned the entire amplifier, retaining the basic layout but greatly reducing noise and distortion. It still sounded like an SA-100, but the way it should have been. The design was pretty much buggered in many ways. I'll be designing new PCBs for the entire thing, same for the SA-220.
These are designed to fail. Not on purpose, but M.E. had no idea what he was doing. He argued and fought with the real design engineers.
I redesigned the entire amplifier, retaining the basic layout but greatly reducing noise and distortion. It still sounded like an SA-100, but the way it should have been. The design was pretty much buggered in many ways. I'll be designing new PCBs for the entire thing, same for the SA-220.
These are designed to fail. Not on purpose, but M.E. had no idea what he was doing. He argued and fought with the real design engineers.
Hi Joe, and all.
This is a very interesting thread, at least for me highly interested in Vbe matching and Hfe matching.
And no, redesign is not the way in my application. There is no other way, but electrolytic caps or a servo i do not want.
Matching two ZTX1053A or using a ZDT1053 ( same but dual ).
About matching measurements, you are aware of the dramatic temperature effect onto VBE.
_Measure at very low Ic to lessen self heating.
_Wear gloves to avoid radiant heat from your hands. ( not hockey players gloves, jeweller gloves ).
_Yes, air flow from a fan is a bad idea.
Statistics.
I understand you see tighter figures with PNP than with NPN.
Do you see a correlation between VBE and Hfe.
Is it high VBE goes with hight Hfe ? Or hight VBE goes with low Hfe ? Or doesn’t depend ?
This is a very interesting thread, at least for me highly interested in Vbe matching and Hfe matching.
And no, redesign is not the way in my application. There is no other way, but electrolytic caps or a servo i do not want.
Matching two ZTX1053A or using a ZDT1053 ( same but dual ).
About matching measurements, you are aware of the dramatic temperature effect onto VBE.
_Measure at very low Ic to lessen self heating.
_Wear gloves to avoid radiant heat from your hands. ( not hockey players gloves, jeweller gloves ).
_Yes, air flow from a fan is a bad idea.
Statistics.
I understand you see tighter figures with PNP than with NPN.
Do you see a correlation between VBE and Hfe.
Is it high VBE goes with hight Hfe ? Or hight VBE goes with low Hfe ? Or doesn’t depend ?
Last edited:
Hi mchambin,
I match beta because it is much more sensitive - and accurate. The jig places the parts in a diff pair configuration as you would use them in your applications. That's why it was designed that way. The two parts are in thermal contact and isolated from ambient and air currents.
I tried many things over the years, and decades ago (late 1980's) I built the first matcher on perf board (wore it out). In practice the matches are far tighter than measuring beta, and I also discovered that previously "matched pairs" weren't well matched just by measuring. The two transistors much be at exactly the same temperature when measured, this jig is by far the easiest and cheapest way to accomplish this. It also is the most accurate.
I match beta because it is much more sensitive - and accurate. The jig places the parts in a diff pair configuration as you would use them in your applications. That's why it was designed that way. The two parts are in thermal contact and isolated from ambient and air currents.
I tried many things over the years, and decades ago (late 1980's) I built the first matcher on perf board (wore it out). In practice the matches are far tighter than measuring beta, and I also discovered that previously "matched pairs" weren't well matched just by measuring. The two transistors much be at exactly the same temperature when measured, this jig is by far the easiest and cheapest way to accomplish this. It also is the most accurate.
Hi anatech,
I fully agree.
In my application the two transistors, at the very same temperature, must match VBE + Rb x Ib.
With the same Ic = 2mA ( this accurately regulated ) and the same bias resistor value Rb = 2 k Ohm.
However not sure, I will not have to change later Ic at 1mA or 4mA and Rb at 1 k Ohm.
This about ZTX1053A that datasheet quotes at Hfe = 400 typical.
I expect sorted transistor pairs within 1mV VBE and 20% Hfe. Is this realistic ?
I fully agree.
In my application the two transistors, at the very same temperature, must match VBE + Rb x Ib.
With the same Ic = 2mA ( this accurately regulated ) and the same bias resistor value Rb = 2 k Ohm.
However not sure, I will not have to change later Ic at 1mA or 4mA and Rb at 1 k Ohm.
This about ZTX1053A that datasheet quotes at Hfe = 400 typical.
I expect sorted transistor pairs within 1mV VBE and 20% Hfe. Is this realistic ?
Who cares about vbe? I don't.
I find that matches have very close vbe, but I don't bother measuring it. So how do you measure vbe keeping the parts at the same temperature? Why not just build the jig.? That eliminates variables. Temperature is a variable and also varies with time.
I find that matches have very close vbe, but I don't bother measuring it. So how do you measure vbe keeping the parts at the same temperature? Why not just build the jig.? That eliminates variables. Temperature is a variable and also varies with time.
Nope.
Match beta. Vbe will match if you are selecting from the same batch once you match beta. Besides, emitter resistors vary by 5% to 10%, and they really only force matching at higher current levels. On that subject, I have seen people replace emitter resistors with dissimilar types that have different temperature coefficients. Oops! ... or should I say ... DUH!
If you're selecting from the same manufacturer, same part number they might match. To be honest, it's been so long since I bothered with Vbe, I measure current sharing once the matched set is installed and running. Never once had any trouble with current sharing.
For diff pairs, beta match using the jig. Equal temperature is critical to judge a match or not. Not unless they are far off.
Match beta. Vbe will match if you are selecting from the same batch once you match beta. Besides, emitter resistors vary by 5% to 10%, and they really only force matching at higher current levels. On that subject, I have seen people replace emitter resistors with dissimilar types that have different temperature coefficients. Oops! ... or should I say ... DUH!
If you're selecting from the same manufacturer, same part number they might match. To be honest, it's been so long since I bothered with Vbe, I measure current sharing once the matched set is installed and running. Never once had any trouble with current sharing.
For diff pairs, beta match using the jig. Equal temperature is critical to judge a match or not. Not unless they are far off.
In my phono preamp (schemo at post #1) the input consists of three parallel transistors sharing the same emitter resistor per polarity; so again in this particular case isn't necessary to have equal Vbe since there isn't an emitter resistor per device to enforce (relatively) equal currents?
Sorry for my insistance, I'm trying to get to the bottom of this, as it has been pointed out to me in another thread that Vbe match was important for this specific part of the preamp, by more than one member.
That being said, your jig is indeed on my list of to-dos👍
Sorry for my insistance, I'm trying to get to the bottom of this, as it has been pointed out to me in another thread that Vbe match was important for this specific part of the preamp, by more than one member.
That being said, your jig is indeed on my list of to-dos👍
In a 66dB DC amplifier a 1mV VBE mismatch means a 2 volt error output.
Measuring: The two transistors are held against each other with a paper clip. The jig has a zero insertion force socket.
Temperature varies with time but is the same for the two transistors that hopefully have the same TempCo.
I forgot to say, I want to firstly match transistor pairs, then the target amplifiers have a trimmer to fine tune the voltage offset, to hopefuly get less than 20mV output error.
Hi mchambin,
In the jig, delta Vbe will tend to show a large mismatch. I have found that beta is more sensitive than Vbe to show a match.
Everything matters, the jig takes Vbe into account. ZIF sockets I find are a pain to use. The turned circular sockets work the best for signal transistors. It is important to use top quality sockets, that should go without saying. I tried cheaper ones (Ebay and AliExpress) and they were simply frustrating. 8 pin dip sockets work well.
Hi SomeJoe,
In your design, things are even more sensitive to device matches. You need to match Vbe and beta as you are trying to make one big transistor for lower noise. Given you are using three parts, you would probably match a pair, then find another to match one existing match you just found. You could make a jig that handles three transistors of course. It would have very limited use for three, but it would measure pairs just fine. You would ignore the third spot when matching pairs. The collector resistors are 100R0 0.1% low tempco parts. I matched those in mine.
In the jig, delta Vbe will tend to show a large mismatch. I have found that beta is more sensitive than Vbe to show a match.
Everything matters, the jig takes Vbe into account. ZIF sockets I find are a pain to use. The turned circular sockets work the best for signal transistors. It is important to use top quality sockets, that should go without saying. I tried cheaper ones (Ebay and AliExpress) and they were simply frustrating. 8 pin dip sockets work well.
Hi SomeJoe,
In your design, things are even more sensitive to device matches. You need to match Vbe and beta as you are trying to make one big transistor for lower noise. Given you are using three parts, you would probably match a pair, then find another to match one existing match you just found. You could make a jig that handles three transistors of course. It would have very limited use for three, but it would measure pairs just fine. You would ignore the third spot when matching pairs. The collector resistors are 100R0 0.1% low tempco parts. I matched those in mine.
I have a single channel built so far and it seems these would help a lot; the original rega design used them.
When setting the two "select on test" resistors (that set the operating conditions of the two voltage gain stages) I observed that current varied a lot with temperature: I ended up setting an initial close adjustment before going to lunch while letting the board warmup then doing the final adjustment when returning.
Some Joe,
Sorry, I am not familiar with your phono preamp topology.
I am working on a microphone preamp that is way more simple. The heart is two very low noise BJTs ( I do not need paralleling ) and two op amps.
Sorry, I am not familiar with your phono preamp topology.
I am working on a microphone preamp that is way more simple. The heart is two very low noise BJTs ( I do not need paralleling ) and two op amps.
lol!
mchambin is right. You have no bias for the input stage, I haven't looked at the rest. Did you substitute BJT transistors for JFets by chance?
mchambin is right. You have no bias for the input stage, I haven't looked at the rest. Did you substitute BJT transistors for JFets by chance?
No, I cloned almost exactly the phono section of my vintage amp as a way to get knowledge in case i need to service it, and have a second MC-capable preamp of proven design.
What do you mean by that?
Ahh, right. They are hard to see in that diagram. Interesting setup.
Hi SomeJoe,
I didn't see the current sources.
I normally see a more definite DC path in the base-emitter circuit. Your base bias current flows between the two sets of input transistor bases and possibly the load (cartridge). The current through the cartridge depends on how much offset there is on the bases. It would be low whatever it is.
What was the vintage amp you took this from? It is an odd design.
Hi SomeJoe,
I didn't see the current sources.
I normally see a more definite DC path in the base-emitter circuit. Your base bias current flows between the two sets of input transistor bases and possibly the load (cartridge). The current through the cartridge depends on how much offset there is on the bases. It would be low whatever it is.
What was the vintage amp you took this from? It is an odd design.