Pete not sure about 1 ohm with mods, but 2 ohm Apogee's panels were claimed to be happily driven in stock form. There was a regulator mod widely offered that was written up in Audio Amateur tho.
Edit> Even the final outputs were run off the regulator. It was offered up in Kit form, so some of the shops could of been doing this install.
Edit> Even the final outputs were run off the regulator. It was offered up in Kit form, so some of the shops could of been doing this install.
Last edited:
Hi Pete,
There was enough confusion about this. To the very best of my knowledge, this amplifier was intended to drive a 4 ohm load - minimum 4 ohms. Some confusion came about from the phrase "2 ohms stable" for example. What this statement means is that the amplifier can drive most 4 ohms loads that may include impedance dips in the 2 ohm region. The amp can not operate continuously at 2 ohms, but brief excursions through a 2 ohm load would not cause it to become unstable and blow up. Adcom was extremely clear on this, we had enough stores connecting very difficult to drive speaker systems on these. Ohm, Apogee and some Infinity speakers were problematic.
Having said that, some times you could run very low resistance loads on these amps as long as the output power was kept way down. There were never any modifications to these amplifiers to handle lower impedance speakers. That would involve a major design change, and more outputs at the very least. Even the power transformers and capacitors would need replacing.
The story on the drivers was an up and down thing. With one set of drivers, clipping the amp without a load would destroy the drivers - then the output stage. The higher voltage drivers solved that, until people started to attempt to drive those stupid low impedance speaker systems. Then the high voltage drivers failed due to excessive current. They couldn't win with the parts selection they had in that case style.
Hi infinia,
I'm sure you're right about the odd number of amps that might behave badly, and how close to the edge you might have to be. Still, if you look at the symasym amp threads, a simple change of output transistor type did cause oscillation. They had plenty of phase margin with the original outputs that were intended. It's a gamble either way.
To my way of thinking, I fix the amp completely. No danger for additional failures when using a variac as long as you know what you are doing.
The regulator mode was available, but I have never seen one - ever. I was asked about it enough, but Adcom seemed to be less than enthusiastic about supplying parts and information.
-Chris
There was enough confusion about this. To the very best of my knowledge, this amplifier was intended to drive a 4 ohm load - minimum 4 ohms. Some confusion came about from the phrase "2 ohms stable" for example. What this statement means is that the amplifier can drive most 4 ohms loads that may include impedance dips in the 2 ohm region. The amp can not operate continuously at 2 ohms, but brief excursions through a 2 ohm load would not cause it to become unstable and blow up. Adcom was extremely clear on this, we had enough stores connecting very difficult to drive speaker systems on these. Ohm, Apogee and some Infinity speakers were problematic.
Having said that, some times you could run very low resistance loads on these amps as long as the output power was kept way down. There were never any modifications to these amplifiers to handle lower impedance speakers. That would involve a major design change, and more outputs at the very least. Even the power transformers and capacitors would need replacing.
Increasing the fuse rating would still void your warranty. This was never an accepted change from Adcom, and UL / CSA would not have allowed this without re-certifying the amplifier.
The story on the drivers was an up and down thing. With one set of drivers, clipping the amp without a load would destroy the drivers - then the output stage. The higher voltage drivers solved that, until people started to attempt to drive those stupid low impedance speaker systems. Then the high voltage drivers failed due to excessive current. They couldn't win with the parts selection they had in that case style.
Hi infinia,
I'm sure you're right about the odd number of amps that might behave badly, and how close to the edge you might have to be. Still, if you look at the symasym amp threads, a simple change of output transistor type did cause oscillation. They had plenty of phase margin with the original outputs that were intended. It's a gamble either way.
To my way of thinking, I fix the amp completely. No danger for additional failures when using a variac as long as you know what you are doing.
The regulator mode was available, but I have never seen one - ever. I was asked about it enough, but Adcom seemed to be less than enthusiastic about supplying parts and information.
That's called rolling the dice. The amp would not be happy, and it certainly would not be sounding at its best. The amp was not rated for a 2 ohm load to begin with.
-Chris
You should read more cautiously, before jumping to these conclusions.
The beta i quoted is (as i wrote it) the MINIMUM beta guaranted
by the manufacturer for the said current, check the datasheet.
What you quote is the TYPICAL beta, which is a mean measured
in a significantly large sampling.
You also seems to not have catched that the curves displayed
by onsemi are not at the same VCE in the two datasheets.
The low voltage beta curve is for a 2V Vce, while the high
voltage are specified at 5V Vce.
OK so if we consider worst case min, it is actually 50 at 1A for the parts I suggest at the current that I believe is more important. You chose the absolute worst case point to make your point which is as I said beyond the spec of the original parts used. I don't play games like this OK, enjoy whichever parts you want to use.
Do you work in the field of semiconductors? Do you know what happens as the process gets better over time?
Yes it was the Apogee that was talked about and I do think it was 2 ohms, some said 1 was even acceptable perhaps as a dip. I do remember the AA mods but I'm sure this shop didn't do those. I've been running mine into 2 ohm loads for over 10 years without a problem. I keep them well ventilated esp. in the summer. I'm going to look into the driver upgrade when I recap them in the near future.
I think that the output stage is over rated and simply will take the 2 ohm loads as long as they are kept cool, there is also the thermal cutout which will shut it down.
I only suggested the cap as a temporary fix on the assumption that substitutions had been made. I don't know how you can look at a schematic and know without any analysis what the gain and phase margins are. Also note that this amp does not only use Miller compensation, there is also C4. I would not change the compensation on an amp that has had an excellent track record such as this one, unless substitutions are made. So, I have to disagree with you unless you.
Pete If you think about it, using a regulator modded into on an existing XFMR supply reduces the available supply voltage by the amount of Vreg dropout + Vripple. So the amp should be able to safely (SOA) drive lower impedance's albeit at reduced Po max.
Last edited:
Yes, I agree and was thinking the same thing. Really, for max power we want lower voltage and higher current from the PS, but moving down in voltage is also the way to go for better SOA.
Agree that the current gain at 1A is quiete relevant in respect of
the expected currents drive.
But then, you have to look at the GBW product of the devices, and
you ll see that the 15030/15031 pair is much more suited for
audio than the high voltage version.
At 1 A , the gbw product of the 15034 has completely collapsed.
Comparison of the curves show that the high voltage devices
have very poor linearity.
For sure, i would myself prefer to use higher voltage devices,
but then, we re making a trade off, exchanging reliabilty for
some performances losses.
If i work in the semi industry? what is the relevance of such
a question in the debated subject ?
Last edited:
I think its just a design philosophy issue, as a semi part is spec'd in the data sheet allows high yields initially, but as the semi process matures the part grouping tends to tighten up away from the WC limits.
Let me ask you wahab, what do you think is the idle bias point collector current for the drivers?
Man, this thread has gone down some interesting paths! I'm still paddling along here, head barely above water. 🙂
In the next hour or two I'm going to pull and test Q1 and Q2 and do some of the tests that were recommended, including checking ECB voltages.
However, I'm not sure exactly how to do that. How much operating voltage should I apply to circuit? Do I run the risk of further damaging components if I run it for any length of time while it's oscillating?
I'm less a noob than when this repair was started, but I'm still a noob...
In the next hour or two I'm going to pull and test Q1 and Q2 and do some of the tests that were recommended, including checking ECB voltages.
However, I'm not sure exactly how to do that. How much operating voltage should I apply to circuit? Do I run the risk of further damaging components if I run it for any length of time while it's oscillating?
I'm less a noob than when this repair was started, but I'm still a noob...
Well, i suppose that you already know the answer, so
i m somewhat just confirming your own thoughts..
Most of the drivers collector current is settled by the power
devices base to emitter resistors, generaly about 100R
in class AB designs, this will amount to about 6mA.
The current drained by the power devices is lower,
in the vicinuty of 1 to 2 mA , depending on the bias,
generaly about 100ma in the said Class AB.
Now, for the amp discussed in this thread, it will be
higher, as the resistor between the two driver s emitters
is only 100R, so just reaching the threshold of conduction
of the power devices will mandate a current of about 12mA
in this resistor.
We must add about 1 to 2 mA for each power devices, so
we ll be close to 20mA, supposing that the amp is class AB.
Of course if the amp is class A , the drivers will see a sizable
increase in collector current at iddle, depending of how deep
the amp is pushed in class A..
I understand there is considerable debate about what to sub for the drivers. Do we at least have a consensus that 15030/31 will work well enough for the circuit to function correctly for testing purposes?
Hearing that they used to cherry pick these at the factory makes me want to buy another Adcom product 🙂
Hearing that they used to cherry pick these at the factory makes me want to buy another Adcom product 🙂
Anatech said MJE15032 and MJE15033 is what he used for factory authorized service.
I would not worry too much about Adcom and cherry picking, mostly early production runs they used what TOS had for drivers. Later on production runs they designed in a different TOS driver to stop all that.
I would not worry too much about Adcom and cherry picking, mostly early production runs they used what TOS had for drivers. Later on production runs they designed in a different TOS driver to stop all that.
Okay, I'll buy some of those. They don't need to be matched b/c the emitter resistors insure current sharing, right?
Why doesn't anyone (at least Newark, Mouser, Digikey) have 68.2 and 682 ohm resistors? They all have 68.1 and 681 ohm. I don't get it.
If I get 1%, are those going to be close enough?
Please try and get the ECB voltages before you remove Q1 and Q2. I'd do it at 40 to 50V on the supply and with no signal, does it oscillate with no signal applied? We might not be able to trust the readings if it is oscillating.
I was going to say, that at least the 15032/33 meet the voltage rating requirement and if that is what the factory chose then that is fine. They are used in an emitter follower that has no voltage gain, and there is a trim pot to roughly match the current gain so no you should not have to match them.
Those resistor values are probably typos, think about the percentage difference and the tolerance you are using, those differences are insignificant.
I'm fairly certain that Nelson Pass has said that he designed a prototype for them, but that they reworked it for production. It was hard to find high voltage parts in those days - it is still an excellent design.
what resistors are you talking about? They have standard values one set for each tolerance
10%
5%
1%
they used to have 2% too
10%
5%
1%
they used to have 2% too
Last edited:
I have never done any analysis on this amp but I would say from experience that your number is in the ball park. I'd say 10 to 20 mA and my point is that these high current devices have all their specification curves ending at 100 mA on the low end. The parts I suggested are spec'd down to 10 mA and 1 mA for the GBP. My point is that with 4/8 ohm loads the drivers will probably never see more than 1A and even then only when the output stage is close to saturation which is pushing things. I might choose a higher current part to improve the design with 2 ohm loads or lower but the output devices do have good beta at high current so this might not even be necessary - more analysis is needed.
If you're going to suggest such high current parts we might as well go with the Fairchild FJP1943/FJP5200, but I would want to prototype it and check stability before I could make a firm suggestion. I'd probably bias it with at least twice the current.
Last edited: