|
Audio Asylum Thread Printer Get a view of an entire thread on one page |
For Sale Ads |
Ha!~ I thought I was finished, but they just keep pullling me back in ... last night, I "completed" the minimal refurb of my Heathkit (aa-121, el-34-based pp amp) by dropping in the much-ballyhooed ssilicon diode to replace the selenium bias rectifier. The voltage was a bit high, so I put a 50-ohm sand resistor in series after the Si diode. It knocked the v down to about 55 (from 57); the schematic says 48 V. The *real* problem, though, is that while the new bias, et al., caps made a real improvement in quietness and hum, apparently the Si diode takes it light-years further, because now there is no output at all, even if I turn the "volume" knobs all the way up (except from the "a" channel TUBES). Bias voltage and balance are steady and right on target (1.5v and 0v, respectively, in both channels). I've got new tubes (Ei) arriving today , but what else could it be? - I can't find a signal loss anywhere, but maybe that's just the 11 pm talking ...
Follow Ups:
Yay! - things are back in swing - all the help was great, but indirectly: while I was testing for the recommended voltages at various points, I discovered that the solder connection to the pwr transformer was bad/intermittent (at the 45VAC point, per the schematic)- I de- and re-soldered it, and twala! clean, fresh (and loud!) sound for two days, nownot too sure if this was a repair, or a repair of a repair, or if there is something else I "fixed" without realizing it, but since all the voltages around the circuit measure up to spec (i.e., within ~10%), the bias holds steady at 0VDC, and there's music(!), I'll live with it for a couple of weeks and decide whether to keep going, or just keep my hands out of further harm's way ...
thanks for the help, everyone! - my wife also sends her regards 8^)
Todd,Not sure if you got your prob figured out as I have not read all of the postings. It seems like w/the change to the diode, the bias voltage is now too negative. This will put the power tubes in the cut off mode and there's no output. When you make the bias more positive, the electron will start to flow and you get current through the tubes. If you make the voltage too positive, too much current will flow and the tube may exceed its power handling ability. Now, I wonder if the bias setting resistors in your amp has drifted significantly. Please check them by unsoldering one end each.
If you look at the link below, specifically at the app data table, you will see the recommended bias voltage based on the plate voltage of your amp. Seems like the voltage should be around -38 V and not -48V that is shown on your schematic. Is there an online link to the schema of the aa-121 on the web somewhere? I have a feeling that even if you put -48V on the grid, it will still be in cut-off mode.
After you get the plate voltage and determine the correct bias voltage, you could figure out the new bias resistor values with the voltage divider formula available on the web. Follow the same topology as the stock resistor setup. Do a search for the formula.
No. If only a few mA of "bias current" exist, you`ll get some output, probably distorted. Something else is wrong here, some mistake is made during the mod.
Thanks for the clarification, Damir. I have not actually play with biasing to the point of cut-off. I was just theorizing that maybe at -55V, there's zero current. Regards.
Todd,First, now don't get mad, but is there a STANDBY switch on your unit? BTW---what exactly, was suppose to measure 48 volts? This is WAY TOO LOW, if this is the power tube's anode plate voltage (usually B+). Usual plate voltages are ten times or more (around 500 volts) for the EL34, in push-pull, AB1 configuration.
Did you check all the B+, C+, D+ voltages after you replaced the bridge? These are the taps that energize the power tubes and the gain stage/phase inverters plates.
The bias supply is usually a separate tap off the power transformer. It has its own rectifier and filter circuit---so it it's possible that your main rectifier is putting out A+, but the other taps (B+, C+, and maybe D+) are NOT connected BUT your bias supply is working.
Result? Measurable A+ voltage. Measurable bias voltage. No sound.
Check the continuity between A+ and B+, if you get no voltage off of B+.
no, not insulted in the least, and no standby switch, eitherthat 48v is from the transformer into the bias supply - that voltage read at 47.8v before I changed the diode out, and since that was the last thing I changed (the diode) before it went silent, I didn't measure voltages throughout, just in the bias circuit ...
Are you shure that you connected the Si-diode right- cathode (dotted end)to the transformer, and anode to condenser. Bias voltage must be negative respect to ground, - 48 V on pin 5, and few hundreds volts on the pin 3 (anode)...
Todd,To follow-up upon what Damir said---the bias voltage MUST be negative with respects to ground. If the measured voltage applied is positive, you risk thermal runaway with the power tubes. This could lead to much more damage to the amplifier.
Be sure that not only the diode, in the bias circuit, has the correct polarity---but also ALL electrolytic capacitors you may have replaced. These also will appear to be placed "backwards" when compared to the filter caps in the main power supply. That is, positive side to ground and negative side to the bias power rail.
I'd pull the power tubes and check the bias supply voltage points. Before venturing further. Next point, -48 VDC is a lot of negative bias voltage for an EL84 amp (if it's in AB1 mode). I've seen this range of negative bias voltage in 6L6 amps.
An EL84 usually has about (-)20-30 VDC applied to the grid, for idle bias. I suppose that your amp may be Class B mode. You should double check those values.
Todd,Re-read your first post. Since the amp is an EL34 and not EL84 based amp---the value of -48 volts (bias) is probably appropriate.
hmm - the IN4004 diode I got at RS doesn't have any marks except the number and a *band* at one end, which I connected to the 50-ohm wirewound, which in turn was connected to the xfrmr output...
O.K. - that`s right, but what`s wrong? Measure all voltages on the output tube sockets:- pins 1 & 8 : cathode and g3, are probably connected to ground with small value resistor (say 10 ohms) for measuring the "standing" current - must measure a few hundreds of mV or so
- pins 2 & 7 heaters - 6,3 V across them
- pin 3 anode - say, 450 V
- pin 4 g2 - connected to UL tap on the OPT, voltage is about Ua (say 450 V)
- pin 5 g1 - "audio input" from phase splitter and "control grid", about - 48 V
- pin 6 is not connected, or is connected like soldering point, for example for screen grid resistor between pin4 & 6Do you have all those voltages? Do you hear the buzz when you touch pin 5 with meter probe? Be carefull, these voltages you must measure with amp "on" - HV present!
thanks for the info. - unfortunately, I'm not in front of the amp right this minute (it'd be a pain to carry it on the train, and since my office is all glass, it'd attract unwelcome attention), but I'll definitely check that - my copy of the schem. is a grainy 3rd-generation photocopy (gee thanks, Heath), which additionally is cut-off at the top and bottom; that, and the fact that I'm learning as I go along (very carefully!) have made for some interesting comments from my basement the last couple of days ...
Todd,Band is positive end. This side should be connected toward the power transformer, when in the bias supply portion of the circuit.
This is something that was drilled into our heads when I went through electronics school (a good few moons ago!)... The arrow always points towards negative material on semi-conductors.... example being transistors. look at an NPN transistor. The arrow points outward from the emitter. The reverse is true for a PNP. I suppose positioning depends on whether the diode is reverse biased or not, but that does not reflect which side of the diode is positive/negative.
the bandend is soldered to the same lug with the xfrmr output - the rest of the caps are soldered correctly - or at least, that's my working assumption, because a) it worked until the new diode went in, and b) they were installed the same way originally (i.e., polarity matches schem. and previous installation)could the higher bias voltage have damaged some other component downstream, causing the signal to break?
Todd,It's possible. Though you probably would not get a negative voltage reading (zero volume, no inputs, no power tubes, and speakers connected) at pin 5, if one of the resistors downstream of the the diode (between the diode and pin 5, of the EL34) was burnt.
Red lead to pin 5 and black lead to ground. Set meter to read 200 VDC range. If you can't get the proper voltage---buy a new diode. Re-insert it and try again. Remember, the voltage should read NEGATIVE.
To test a diode, it should let voltage through in one direction. But, not the other. Ohm (DC) setting on the DVM can quickly test this. But, this is a crude measurement and can't tell you the degree of potential thermal damage. Better to buy another diode.
Measure the voltage on pin 3, without the EL34 tubes inserted. Should be around 500 VDC (positive). Set the DVM to 1000 volts VDC range.
Red lead to pin 3. Black lead to ground.
If pin 5 reads -48 volts VDC and pin 3 reads around 500 volts VDC---and you still can't get any sound from the speakers (after you've turned the amp off and inserted the EL34's)---check for lit plates on the tubes.
(The reason I suggest not putting the EL34 into the amp, while testing for voltages---is to prevent any further damage to the tubes. Your reading maybe a tad higher, since there will NOT be the load of the tubes to drop the readings).
Also, check for 6.3 volts VAC across pins 2 and 7. Maybe, you have a bad connection to the heater pins. If okay, plug in the EL34's.
If the heaters are fine. Anode voltage is fine. Bias voltage is in proper range. And still no noise from the speakers.
Be sure all tubes light up. Check the heaters, again. 12A(_)7 should measure 12.6 volts across pins 4 and 5. 6SN7 types measure 6.3 volts across pins 7 & 8. 6FQ7 measure 6.3 volts across pins 4 & 5.If for some reason the current draw was too much in the power stage---a fuse could have blown. Probably need to hunt around for any blown out fuses. These can be hidden on a circuit board (on the top or bottom of the board).
It could also be a bad coupling cap(s) from the phase inverter or gain stage.
Check for continuity with the amp off and DVM set to Ohm's (DC) in the Meg ohm range. Could be a bad solder. Bad cap. Bad resistor in the RC circuit of the gain stage or PI. OR a bad tube in these sections.
One other option you don't even want to think about is a blown OPT. There are ways to test the continuity (Ohm meter, again with amp off). But, this is crude and can really determine only breaks in continuity or shorts.
Give the TA board the results of some of these tests & maybe we can help narrow down the problem area.
Good hunting.
Todd,ONe other thing, many amps have a screen grid resistor which acts as kind of a fuse---to prevent excess current flow, within the tube (and potential thermal runaway).
In Marshall amps a 250-1000 ohm, 5 watt resistor is tied to pin 4 of the EL34's. Check this resistor for its measured value. And any signs of the "burnt toast" effect.
as this gives me plenty to do, and I can avoid the usual tour-guide stuff ("no, really, I'd love to see Stone Mountain for the 150th time ...")thanks for all the help, and I will post on this once I get it fixed (or come to another major deadzone) ...
This post is made possible by the generous support of people like you and our sponsors: