"SOC1A" this is what I can read on the power amp fuses
The main fuse was 0.5A not 5A, my bad I didn't check it earlier... I just checked that it wasn't broke and I wrongly assumed it was the right one without read, but honestly who can imagine someone has used a 0.5A for a 6A
Regarding the heat output, I seem to recall from Vintage Knob that these amps were biased quite highly, in fact they even made one in the same casing that was class A.
I normally bias my amps at 30mA idle, and this is plenty. 22mA is close enough
With adequate ventillation around the unit, you should be OK.
I normally bias my amps at 30mA idle, and this is plenty. 22mA is close enough
With adequate ventillation around the unit, you should be OK.
Regarding the heat output, I seem to recall from Vintage Knob that these amps were biased quite highly, in fact they even made one in the same casing that was class A.
I normally bias my amps at 30mA idle, and this is plenty. 22mA is close enough
Thanks Jaycee. I looked back to the work Doug Self did on output stage distortion and optimum bias currents. I wanted to play it safe as well by not overbiasing.
For 0.47 (0.5) ohm emitter resistors a value of around 27 millivolts was mentioned giving around 58 milliamps. Drop the emitter resistors to 0.1 ohm and the current needs to be over 200ma for optimum performance. Swings and roundabouts... the 0.5 ohms allow lower currents but at the expense of efficiency.
gde3... get it all back working with the bulb tester again. Lets be 100% sure all is OK
One other thing you can check is this (because I want to be absolutely sure everything is good to go)
Using the bulb tester and with VR2 on minimum again, measure and record the DC voltage across R24
Be extremely careful measuring as one slip would be a disaster
Now adjust VR2 as before to give around 22 millivolts between 9 and 13
Now measure and record the voltage across R24 again. It should be the same. If it is you can carry on with testing on full mains voltage. If it isn't then wait.
Return VR2 back to minimum and switch off.
Reason for doing that is to be absolutely sure that the rails aren't dropping to much under test with the bulb and causing the current in TR11 to fall.
Good luck... I'll look in later.
One other thing you can check is this (because I want to be absolutely sure everything is good to go
)Using the bulb tester and with VR2 on minimum again, measure and record the DC voltage across R24
Be extremely careful measuring as one slip would be a disaster
Now adjust VR2 as before to give around 22 millivolts between 9 and 13
Now measure and record the voltage across R24 again. It should be the same. If it is you can carry on with testing on full mains voltage. If it isn't then wait.
Return VR2 back to minimum and switch off.
Reason for doing that is to be absolutely sure that the rails aren't dropping to much under test with the bulb and causing the current in TR11 to fall.
Good luck... I'll look in later.
Attachments
gde3... get it all back working with the bulb tester again. Lets be 100% sure all is OK
One other thing you can check is this (because I want to be absolutely sure everything is good to go
Using the bulb tester and with VR2 on minimum again, measure and record the DC voltage across R24
Be extremely careful measuring as one slip would be a disaster
Now adjust VR2 as before to give around 22 millivolts between 9 and 13
Now measure and record the voltage across R24 again. It should be the same. If it is you can carry on with testing on full mains voltage. If it isn't then wait.
Return VR2 back to minimum and switch off.
Reason for doing that is to be absolutely sure that the rails aren't dropping to much under test with the bulb and causing the current in TR11 to fall.
Good luck... I'll look in later.
2.613 DCV on R24 with VR2 at min an R1 at 0
With bias at 22mv DCV on R24 is 2.315
Should I continue the test without the bulb?
Wonderful... the excitement is to much... I'll have to go and have something to eat
Don't rush... you have to be sure the bias on this channel is stable and not prone to runaway with temperature.
So use that channel, turn it up and get the heatsink hot and then turn the sound down and recheck the bias. We have to be sure
When we are sure its OK we work on the other channel.
The other channel has to be set up in the same way but aren't we using the original drivers on that one ? So you don't alter any resistor values on that one.
Remember to start with VR2 on that channel on minimum too.
Don't rush... you have to be sure the bias on this channel is stable and not prone to runaway with temperature.
So use that channel, turn it up and get the heatsink hot and then turn the sound down and recheck the bias. We have to be sure
When we are sure its OK we work on the other channel.
The other channel has to be set up in the same way but aren't we using the original drivers on that one ? So you don't alter any resistor values on that one.
Remember to start with VR2 on that channel on minimum too.
Don't rush... you have to be sure the bias on this channel is stable and not prone to runaway with temperature.
So use that channel, turn it up and get the heatsink hot and then turn the sound down and recheck the bias. We have to be sure
Here's a handy trick: use a hair drier and blow the hot air over the heatsink (use the maximum setting). The air isn't hot enough to do damage but will substantially heat up the heatsink in a relatively short time without the need to have an input signal. Monitor the voltage across the emitter resistors. It will not be exactly constant but at no time should it start creeping up faster and faster!
Yes, that's a good tip to use a hairdryer. If you do, heat the metalwork evenly.
To be continued... tomorrow
After setting both VR1-2 I tested both channels separately (without the bulb) everything fine
Tomorrow I will finish the assembly and I will try the amp with good speakers, stay tuned
Just a couple of questions (sorry if some are really noobish)
- Using only 2 pairs of transistors instead of 3 can generate unbalance between the channels?
- I'm not interested to run this amp in class A, but if you guys have more informations on this topic I will be happy to read
- The big caps are working right for the moment. Substitute them is something I should worry about in the future?
- The amp has a subsonic filter (should work up to 8hz), on the manual it say to always leave it on since without there could be distorsion. Can someone please explain me a little further?
- The preamp has six AC output. Is it better that I connect the amp to this power source or I keep amp and preamp separated? (in the first option a gnd cable is also connected between the two)
3 of this AC outlet are unswitched and rated up to 350W, 3 are switched 800W
- Using only 2 pairs of transistors instead of 3 can generate unbalance between the channels?
- I'm not interested to run this amp in class A, but if you guys have more informations on this topic I will be happy to read
- The big caps are working right for the moment. Substitute them is something I should worry about in the future?
- The amp has a subsonic filter (should work up to 8hz), on the manual it say to always leave it on since without there could be distorsion. Can someone please explain me a little further?
- The preamp has six AC output. Is it better that I connect the amp to this power source or I keep amp and preamp separated? (in the first option a gnd cable is also connected between the two)
3 of this AC outlet are unswitched and rated up to 350W, 3 are switched 800W
I would probably say fit 3 pairs if that was what the amp was designed for, but other than having less current capacity with only 2 pairs, there should be no sonic difference.
The Pioneer M4 was class A. I am not sure if it's the same circuit with the bias cranked up high or not. I saw it here: Pioneer Exclusive M4 on thevintageknob.org
You may find the big capacitors are OK. The telltale sign of them being near the end of their life is hum and noise on the amplifier, caused by ripple voltage the capacitors are failing to smooth.
Subsonic filter - try with and without, use whichever sounds best to you
It is not necessary to use the preamp's AC outputs, I guess there is just the convenience of having only one power switch.
The Pioneer M4 was class A. I am not sure if it's the same circuit with the bias cranked up high or not. I saw it here: Pioneer Exclusive M4 on thevintageknob.org
You may find the big capacitors are OK. The telltale sign of them being near the end of their life is hum and noise on the amplifier, caused by ripple voltage the capacitors are failing to smooth.
Subsonic filter - try with and without, use whichever sounds best to you
It is not necessary to use the preamp's AC outputs, I guess there is just the convenience of having only one power switch.
No, the amp corrects for this. However, using 2 pairs instead of 3 results in a lower maximum permissible current from the amp, which in turn means you must use a higher speaker impedance with it if you want it to work within it's specifications, and in some cases it's necessary to have the amp work reliably.
Why is this important? Although speakers are declared with a NOMINAL impedance, this is actually variable, most notably with frequency. If your amplifier is specified for an 8 ohm SPEAKER load, it's good design practice for it to work reliably on a 4 ohm resistive load (which is the easyest to drive and constant). As you can expect, if it's specified for a 4 ohm load, it is likely designed to work reliably on a 2 ohm resistive load. Now, simple ohm's law will tell you what the peak current is into this sort of load - the power supply voltage / 2 - in your case about 30A! Mind you, this is a peak current and is distributed among three pairs, and also one half of the pair works for one half of each cycle (either positive or negative) so each transistor only sees a much lower average - but as you may notice, with two pairs, this is at the maximum limit per device (15A). There has to be some margin, for various reasons I won't go into here, which are commonly referred to by the concept of transistor 'safe operating area' or SOA, which specifies what combinations of currents and voltages are safe for the transistor.
Also, there is the question of performance. With transistors, current gain falls off over a certain current, and the peak current mentioned above will be well over this limit for only two pairs. As a result the distortion generated in the output stage will be higher than expected.
Keep in mind that I have used 'worst case' scenarios for both considerations mentioned above. However, it's not altogether unlikely for a pretty regular speaker to have impedance dips and peaks which may at some frequency indeed be close to the worst case scenario - and when the source signal 'hits' it, at the right loudness, you might get unexpected audible effects. It canbe said that it's a measure of the amp's quality how independant it's measured and audible performance is with respect to the load, and in your case the people who designed it deliberately 'over-designed' it to insure good performance.
Your amp cannot reliably work in class A above a rather limited power, about 1W or so. Let me explain this a bit. Class A is commonly defined as a condition where neither the top or the bottom half of the output stage ever completely turns off throughout a full cycle of a signal. The way this is done is usually by increasing the bias current of the output stage to a point where it's equal to or larger than the half the maximum load current at rated load and power. The 'half' bit is confusing somewhat, but the idea is that when the signal is zero, then half the maximum current flows from + to - power supply. When the output is fully positive or negative, the full current flows from either the positive or the negative supply intop the load, but no current (Zero) flows from the oposite power supply to the load. For all other output cases the situation 'slides' between those two - when current increased from the + supply to the load, the current from the - supply is decreased by exactly the same amount - this is why mid position ends up being half from one, half from the other, which cancels them out giving you a zero output.
Now, the problem here is primairly heat. This sort of amplifier actually always produces the same power but when it's not going out to the load, it's turning into heat. It takes a HUGE heatsink to get that much heat away from the transistors. In fact, because of the SOA thng mentioned above, that might not be the only consideration - it might (and typically does) turn out that having the full supply voltage and half the current in the output stage is far beyond their SOA (as mentioned above) and even the ability to cool themselves as such, so many more pairs are needed. Or, as is the case more commonly, the same heatsink and transistors can only make a small fraction of power in class A compared to class AB, which is normally done by using a much reduced power supply voltage for the output stage. At some point you might want to look at a Pioneer M22 which uses exactly that approach, and 2 pairs of transistors on a huge heatsink - but only 30W per channel.
All that being said, your amp works in class AB, so there is something A in AB. It has a relatively small idle current, which passes from V+ to V- in the output stage at output=0. This is because the transistors are kept slightly active, to avoid the situation where none of them would work at all if the signal was below a certain treshold - at this point you would get nothing out at all, and for slightly larger signals the output would be grossly distorted. For very high outputs the 'missing' part around zero would be quite small so it would not be as gross but the distortion would be clearly audible for more silent passages of music. In order to avoid this, there is the idle current, a 'default' of sorts which makes both transistors work a bit around zero, so there is no 'missing part' - at any point at least one is working (compared to class A where at any point BOTH are working), and for small signals around zero, both are working - in this area the amp behaves just like it was class A, but only as long as the output current is up to twice the idle current - so, because the idle current is far below half of the maximum, a class AB amp actually works in class A but only up to some small fraction of it's maximum power. Although this does not sound like even close to class A, in actuality, a music signal has very high peak power but very low average power, and for most listening levels will be well below 1W (the ear is logarithmic in it's perception of loudness, so 10W is only about twice as lous as 1W - conseqeuence being you need a LOT more power for just a bit more loudness!!!) - and so, for most of the time your amp will indeed work in class A.
At this point you might think there is something to gain by increasing the idle current, and thus getting more power in class A (keeping in mind this also produces more heat, so there is a limit). Well, there is a problem with this. Namely, the idle current turnd on both top and bottom transistors in the output but right around the point of turn-on, the transistors are very nonlinear. There is a relatively small range of idle currents that makes it possible to overlap the nonlinearity of the top and bottom half of the output stage so that much of it cancels out, and, no surprise, the ide current you have set is right inside those limits. BElow it, you will get vastly increased distortion for small signals, and above it you get increased distortion for intermediate signals, right over the limit of class A, but close to the average listening level - definitely not something you want.
Manufacturers of this level of components chose the best they could get at the time, so it's not unusual to have these big caps within spec even after 30 years. However, there is one point you should think about. If the actual operating voltage is very close to the maximum (specified on the caps), say about over 85% of maximum, they might have lost quite a bit of their rated capacitance. It should be noted that this class of component will be over-designed in this area as well, however, the lack of storage capacitance might become apparent at higher power and lower impedance loads (usually when measuring you can't get the full rated power at the specified distortion level).
Back then filter caps were specified as +50/-10% or +30/-5% tolerance on capacitance. Voltages close to the maximum spec will result in reduction of the capacitance much quicker, but it does take some time to 'eat up' the usual 30% more you got on new caps. BUT - and here's where the problem is, this was when the mains voltage was 220V. As it's now been 230 for 2 decades, some amps that have been in operation that long will have spent more than half their lifetime with a voltage on their caps quite a bit closer to the maximum than originally anticipated. You mentioned that this amp has been sitting broken for a long while so you may not have that problem now, however, the caps are old, and now it's working on a higher mains voltage, so new (with higher voltage rating!) caps might be something you want to think about at some point, especially since you have put so much work in the amp already. That being said, as far as i can see it's not a terrible job to replace them afterwards, so I would not consider it a priority.
Because the amp is DC throughout, an input signal with subsonics will end up amplified and fed to the speaker. Subsonic signals can gobble up a lot of power without being immediately noticed because they are not audible. Also, they can damage speakers because of this - pushing the voolume up will increase the power contained in the subsonic part to dangerous levels without being heard immediately. Even if the speaker can take it, the subsonic part might end up driving the amp to clipping well before it's expected. To prevent this, there is a subsonic filter. Now, wether you use it or not depends on expecting this sort of problem. Subsonics normally come from turntables, and are generated by warped records, or low frequency rumble transmitted from the surroundings (for instance footsteps). They also sometimes appear in orchestral music, from pipe organs, and very often in electronic music. At low volume it will hardly make a difference. If you suspect a problem (look at the membranes of your woofers - if they 'swing' back and forth slowly, you have a problem!) you can switch the sub filter on.
It's a sort of integrated power strip. It's designed to enable you to use one switch for everything. However, I would advise against that for the preamp, and the reason is for proper power-on sequence. The power amp is always switched on LAST and off FIRST. The reason for this is that yiu don't want any power on or off transients from the sources or the preamp to be amplified by the power amp, which could potentially damage the amp and speakers. Therefore you can use the unswitched outlets if you need to but do use the power switches on your other gear properly, rather than relying on the power amp power switch to switch everything on at once, expecting no transients. (It is true that the turn-on delay will keep speakers switched off during any power-on transients but even though that means the speakers will not see them, the amp will, which may not be a good thing, and then there is still the turn-off transient to deal with).
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.