Heathkit AA-151

[Linn lover]

DR,

I will get voltages measured in detail this weekend and get back to you with all the measurements I take.

[dcgillespie]

Linn -- The values used in the original design for the 6AN8 are quite appropriate as is. However, some of the actual components used are rather famous for drifting very high in value -- sometimes by many times their original value over time. In particular, the 680K screen grid resistor, and the 220K plate load resistor (both for the pentode stage) are particularly guilty of this, so I would therefore check their values very carefully if they are still the original components installed.

I have no doubt you will ultimately get the amplifier performing as it was originally intended to perform.

Dave

[Linn lover]

Dave,

When I had the tubes out, with resistors still in place, I did some measurements.

220K was measuring over 250k (in circuit, no tubes present)

680k was measuring over 720k (in circuit, no tubes present)

I will remove them and measure out of circuit. If I replace them, will 2 watt resistors be sufficient? I have some of those lying around.

From this point on, it can only get better!

[battradio]

Dave,

When I had the tubes out, with resistors still in place, I did some measurements.

220K was measuring over 250k (in circuit, no tubes present)

680k was measuring over 720k (in circuit, no tubes present)

I will remove them and measure out of circuit. If I replace them, will 2 watt resistors be sufficient? I have some of those lying around.

From this point on, it can only get better!

For fruture referance if a resistor mesures high in circuit it will always mesure high out of circuit , If it mesures low in circuit it needs to be checked with at least one end loose .

[dcgillespie]

The wattage rating of the resistors is a bit of a game. The original units were 1/2 watt, and completely adequate for the job electrically. However, increasing the wattage rating usually reduces noise -- but if you increase the wattage rating too far, the physical size of the component can cause other issues like capacitive coupling to adjacent components. I would replace them with 1 watt components. A 1 watt resistor today is almost no bigger physically than a 1/2 watt resistor from yesterday, so going with a 1 watt unit will eliminate any issues due to physical size, and also give you the added benefit of lower noise.

Dave

[Linn lover]

Thanks Dave.

As I kept remove components and replacing them with newer stock, I have noticed that more and more resistors are out of tolerance. The 150k are now measuring, out of circuit, over 170k, so I will be replacing those too.

The four 470K at the input of the EL84 (6BQ5) are also reading way too high, like 520K.

I measure all values prior to installing them in the circuit. All is good to date, but more work remains.

I will install the 47 ohm resistor (10 watts), 30 uf (450V) capacitor at the output of pin #8 of the tube rectifier. I will do this until the choke comes in to reduce voltage inside the unit. I found a good plate for the components today, but have to solder them in.

At this point and time, patience is a virtue. I cannot wait to see if all these changes will yield good results. One step at a time.

Appreciate your technical know how. I will let you know when I fire the AA-151 up again.

[Linn lover]

DC & Dr,

I have changed many resistors. I have not changed R24, R52 (22k), R25 R53, R26, R54 (36k) yet. They measure about 25k and 38k respectively.

I have played around with the power supply. I even tapped the 275 volt source to see if that would help the 6AN8s. The voltage after that change measured was too low (for example, 159 and 147 volts at pin #1 of the 6AN8, where you should have 190 volts).

I reverted back to It's original tap (which should be 285V), but I needed to get the voltage down since it was measuring high. I installed a 30uf (450V) and 47 ohm resistor (10 watt) prior to the power supply. It helped a bit by dropping 5 volts from 399 to 395. Not enough. So I increased the resistance until I saw something acceptable: now sitting with the cap (30 uf - 450V) and a 220 ohm resistor.

It's still not perfect, but it sounds much better. The voltage report has improved, but I am at a lost why there are still discrepancies at the 6AN8. At least when I drop voltages, things improve. I will change the 22k and 36k resistors in the hopes that a problem lies there that I cannot see.

Here is the voltage report:

Power supply source: 366 v (used to be 399V)
Power supply after the 4.7k resistor:316V (used to be around 335 I think) * Maybe increase the 4.7k to a higher value? I don't seem to be dropping enough across that resistor.

EL 84:
Pin # 9 = 363v
Pin # 7 = 359 v

6AN8

Pin #1 250v and 225v
Pin #2 123 and 124
Pin #3 112 and 112
Pin #4 54 and 54
Pin #5 54 and 54
Pin #6 51 and 88
Pin #7 49 and 40
Pin #8 0v and 0v (some small voltages, but like 200 mv at best)
Pin #9 1.5 and 1.4

The supply voltage is 316 to the 6AN8. I need to reduce that to 285 as per the plan.

Comments gents?

[dcgillespie]

Since you are using this amp as a power amp only, I am assuming you have removed all irrelevant tubes. If so, this is why (in part) your down stream B+ voltages are all too high. To correct, you will need to adjust the 4.7K B+ dropping resistor appropriately to compensate. I would start with double the value (10K) and see if that doesn't get you closer to the mark.

With the 36K resistors, the issue is not so much one of absolute value, but that each are matched to each other within a channel. The better these resistors are balanced, the better the balance will be in drive to the output stage. Try to achieve a 1% balance between these two resistors (in each channel), while keeping the absolute value within 5%.

The 220 ohm resistor in the power supply will definite hurt power output, since this is a class AB1 amplifier. I would not worry so much about absolute voltage, as I would about the dissipation of the output tubes. We need to know what the cathode voltage is, and what the actual value of the 100 ohm cathode resistor is. From that we can derive current draw per tube (assuming the tubes are relatively matched), and then this with the plate voltage will let us determine dissipation. This should be measured with the original power supply configuration in place. If that exercise indicates that the B+ needs to be dropped -- yet the heater voltage is in line -- then the best way to do that is to use a string of (say) 4 or 5 6 volt zener diodes in series that add up to the required drop. These should be 5 watt devices, but will each dissipate less than a watt for good long life dependability in this scenario. If the heater voltage is higher than +5%, then the better way is to drop the line voltage with a bucking transformer, which will drop all voltages within the amplifier.

The advantage of the zener approach (if used) is that the B+ will not drop as the output stages draw more current with power output, where as it will drop significantly with the 220 ohm resistor approach.

When to B+ supply to the 6AN8s is correct, and the resistive values associated with these stages are also correct, then the voltages at the tube elements will be what they will be. This is because even small manufacturing tolerances in the screen grid of the pentode section of this tube will produce wild voltage changes at the screen and plate of the pentode section -- and therefore at the triode section as well due to the direct coupling used between these stages.

Great work!

Dave

[Linn lover]

Dave,

Yes, the amp is indeed being used as a power amp.

Prior to the power supply mods, the plates were at 388 v, the cathode resistor is 120 ohm and the voltage was 17 volts: hence the current would be 35.4 ma per tube (they are not matched as I am using the old ones). Power per tube is approx 13.8 watts.

I have a matched set that I will plug in when I have everything under control.

For the power supply, the choke is on order which would replace the series resistors...or in addition to it.

With the mods to the power supply (the 220 ohm and 30 uf cap), the plates are at 363 v, the cathode resistor is 120 ohm and the measured voltage is 16 volts. Current per tube would be 33.33 ma and power per tube about 12 watts. Correct?

Next step: work on the power supply first line resistor...the 4.7K...replace with 10K or so.

My 6AN8: they tested ok on a tube tester, but, they are the original tubes. Should I replace them with new ones after I replace the resistors. That would make things more accurate?

One step at a time. I am going to try the amp with the 220ohm in to see what it sounds like with my Reference 3A Decapo MM (one original, one homemade) - only after I change the remaining resistors and get the power supply voltages under control.

OK - Time to change a few more resistors. I am starting to feel that I know what I am doing now...I think they call that learning!

Cheers Dave!

[Linn lover]

Dave,

According to my calculations, I need to drop 81 volts (366-81 = 285) across the first resistor in the power supply.

If I calculate the current across the measured value of the resistor, total current is approximately 9.8 ma.

81 volts and 9.8 ma yields a resistor value of approx 8.2 k ohm. I'll put in a 2 watt resistor to be safe...lots of margin there.

[DeathRex]

6AN8

Pin #1 250v and 225v
Pin #2 123 and 124
Pin #3 112 and 112
Pin #4 54 and 54
Pin #5 54 and 54
Pin #6 51 and 88
Pin #7 49 and 40
Pin #8 0v and 0v (some small voltages, but like 200 mv at best)
Pin #9 1.5 and 1.4

Pins 7, 8, and 9 are good on both. Pin 6 is good on one, but little low on the other. Pins 4 and 5 don't count (filament). Pins, 1, 2, and 3 are all high with pin 2 (grid) much higher than pin 3 (cathode). That's almost impossible. Also pin 2 has to be the same voltage as 6, and it's not, that's impossible, except if the triode section was drawing grid current. I've only seen a EL84 draw grid current. It's kinda strange that both 6AN8s are doing the same thing. With these reading, I'm betting both 6AN8s are very bad. Can you test them?

I ended up ripping out the 6AN8s out of mine and cloning the Shannon's ST-35.

[dcgillespie]

Linn -- Your calculations are correct for the 4.7K resistor. Also consider however, that with reduced B+ to the 6AN8 via a bigger dropping resistor, the current draw will be slightly less than it is drawing now. That will work to move the required value even higher for the target voltage drop. That's why I suggested 10K to account for that, but either value will get you comfortably in the ball park, and then you can tweak from there.

As for output tube dissipation, two points to consider:

The voltage in question (in this case) is not just the plate voltage, but the plate voltage minus the voltage at the cathode. Therefore, using the original figures, 388 plate volts minus 17 cathode volts equals 371 volts across the cathode and plate elements. That is the voltage used to determine dissipation.

2. 17 volts across a 120 ohm cathode resistor serving four tubes does equate to 35.4 ma per tube (assuming they are matched), except that this is total current draw for each tube, including that of the screen. For a true pentode tube like a EL84, you would need to subtract about 3.5 ma from this figure to arrive at the current flow actually reaching the plate.

Using the corrected figures then of 371 volts and ~ 32 ma then, shows the plate to actually dissipate about 11.9 watts.

Against the conservative Design Center rating of 12 watts for the tube (which was the rating system in place when this tube was first made available), the tube is basically maxed out dissipation wise.

Against the more modern Design Maximum rating for the same tube, it is actually running at just under 83% of ratings.

When you do not take the rating system into account, you can come to the wrong conclusions real quick. Design Center ratings took into account normal variances that could take place both in the tube, and the circuit it was installed in. With a Design Center rating of 12 watts, that is why the engineers of the day could run the tube right to the max, and it was still considered as good design practice. When folks today see that it's running at the published max but don't understand the rating system used, you then see all the disparaging remarks come out about the fact they only did it because of plentiful, cheap tubes, a power output race, chasing the Nth degree performance, etc. -- implying the engineers didn't really care about the "abuse" their design put the tube through. In reality, they could do it, because the design rating system for that rating made allowances so that they could, and do so safely.

The modern Design Maximum rating system does not allow for any variance in the circuit, and so tube ratings published under that system should never be exceeded under any conditions of use, and if possible operated at about 80% of those maximums.

I hope this helps!

Dave

[Linn lover]

DR,

I will double the check the circuit and make sure all is connected properly and measure again!

I have replaced all 22k and 36k resistors tonight around the 6AN8s.

Tomorrow I will drop the power supply voltage to a target of 285 volts.

Then I will re-measure. Don't know what to tell you. I am sharing the information I am measuring.

Maybe I need new tubes!

Back to you tomorrow evening sometime.

[dcgillespie]

BTW -- Heath lists the voltage supplying the 6AN8s at 305 vdc on their schematic. But maybe you didn't want to hear that!

Dave

[Linn lover]

Dave,

I have installed the 22k and 36k resistors tonight. The ones I had were all very close in value and identical except for one resistor was off by 150 ohm (2% tolerance rating).

All resistors I removed were indeed matched, but at 25k and 38k respectively.

I have removed the 4.7K 2 watt power supply resistor. I will pick one up tomorrow....a 10k and a 8.2k to see where that gets me. As per your information, maybe 10k to start is best.

With the 220ohm power supply resistor in place, the plates are at 363 volt and the cathode resistor is sitting at 16 volts. With 3.5 ma removed, that leaves approx 29.83 ma per tube. Voltage difference is now 347 volts, and the power output would now be 10.3 watts.

Heathkit rated the amp at 14 watts per channel! See link attached:
http://www.heathkit-museum.com/hifi/hvmaa-151.shtml

I guess they ran it hot!

Over and out for tonight. Thanks again for the education!

Jules (AKA - Linn lover).