Federal AM864 troubleshooting help

Ive been working on a federal AM864 all day today and have hit the wall. check it:

http://www.montagar.com/~patj/am-864u-sch.gif

This damn thing is so simple Im pissed I cant track this down.

I got:

1. 60 hz ish Hum.

2. Lotsa hiss.

3. No compression.

4. waaaaaay too much output level.

So far Ive:

replaced the V4 with a known working tube. Unit behaves the same with the tube installed and with it pulled entirely- no compression

put it a known working 5y3, get the same hum

Replaced c2, c3, c5, c6

replaced threshold pot.

It appears the previous owner bypassed C4 with a 47uf/450 electro on either side of L1 with the negative side of the cap to ground.

Seems that as soon as I open the input attenuator the littlest bit, the unit goes to full output gain. Im assuming this is where the hiss is coming from.

get the following voltages on V4:
7- 6.8 vac
6- 183 v (manual shows 140 normal)
5- -.54v
4- -.54 (manual shows -1.42 normal)
3- 2.5v

Any thoughts? Is there something obvious to check? Is that 1 volt discrepency on pins 4/5 ruining my day there? Could R7 and R6 be throwing off the voltages on pin 6 and 5/4? I have a feeling the voltage differences on those pins are a symptom, not the problem, but I really dont understand enough to draw a conclusion either way.

dave

Did you check V1 and V2 to see if the fil is good, maybe one of the input tubes are bad. If one is bad the circuit might not work correctly 60 hz ish Hum and hiss.

The cap can C4 was bypassed, indicating that it went bad at some time. Remove C4 entirely from the circuit; leave the can bolted on the chassis for looks, if you like. Leave the "new" filter caps in-circuit, of course.

You didn't mention anything about adjusting R8. Did you try it? That control varies the plate current (and amplification) of that tube as well as affecting the action of the internal signal-rectifying diodes on pins 4 and 5. If that pot goes flaky, you won't get any compression.

Of course, since this is somebody's re-draw of the original schematic, accuracy is always suspect. There used to be an original schematic online, but it seems to have vanished.

I adjusted r8 for 2.5v as the manual suggests, however on my working unit I like the sound of the unit at around 1V. Anyhow, the pot swings up to about 11V on pins 4/5, so I am assuming the pot is working fine, I couldnt find dead spots. Adjustment had no effect on the problem.

The schem I linked matches the schem in the manual except mine shows r3 and r4 as 12K.

I swapped the input tubes on gus's suggestion. Got better performance out of the unit, but all the same problems. I measured some voltages-

V1/V2
pin 4 : -.7 (manual shows -1.3v)
pin 6 : 11.3v (manual shows 118v, dont know if thats a typo)
pin 7 : 6.6 vac
pin 8 : 173 v (manual shows 155v)

v3
pin2 : 181 v
pin 3 : 3.9 v
pin 5 : 180v
pin 6 : 3.9 v
pin 8 : 6.7 vac

v3 voltages are basically spot on

v5
pin 2: 192v (shows 208v)
pin 4: 189vac (shows 175v)
pin 6: 190vac (shows 175v)
pin 8: 191v (shows 208v)


The sound of the hum I am guessing is coming from the compression circuit. With the threshold turned all the way up, the hum is minimal, however when the threshold is lowered and the unit goes into compression, the hum becomes very pronounced.

Additionally, how does the meter calibration in this design work? There are no obviously wrong values in the box compared to the schematic but Im getting the following:

With the meter in output metering, with 0dB on the meter, I measure -4dB at the output. With the meter in attenuation metering, I get 0dB at the output. Is it normal for a meter to cause a -4dB insertion loss?? With the better set of input tubes the most amount of attenuation I can get out of this unit at unity is -4dB. Also, in order to get the GR meter to sit at 0 (instead of off scale) I needed to add resistance in parallel at r2. With that resistor in or out it didnt effect the output level or sound of the unit, only where the needle sat while in GR metering mode.

I cant believe Ive been at this for the whole freaking day, totally got sucked into this one. Damn.

dave

Connect a voltmeter across the grids (pin 4) of V1 and V2. See if there's any imbalance in DC voltage on the grids as the unit goes into compression. With the voltmeter connected across the two grids, you'll see zero volts when the grids are receiving the same bias voltage. A significant offset between the grids points to a problem. The fact that the hum gets worse as the unit goes into compression leads me to believe there's an imbalance there. Do the same across the plates (pin 8).

And I still strongly recommend taking the defective cap can out of the circuit. Have you put a scope on the power supply to make sure that some of your problem isn't coming from there?

ok, between pin 4's, at no compression I get -0.4mv and at max compression I get -0.1mv.

between pin 8's, at no compression I get -1v and at max compression I get -0.73V.

Im assuming the difference between pin 4 is irrelevant, is one volt on pin 8 something to be concerned about?

dave

Nah, sounds like you're pretty well-balanced there.

I'll turn this over in my head tomorrow when I'm more alert. If you haven't fixed it already by then, I'll think of more things to check. It's funny, troubleshooting by proxy is much harder than doing it hands-on.

I have one of these units too. I am having problems too..
my unit fires up and passes audio with some hiss but no real 60hz problem.
the needle goes all the way to the end of the scale and ther is no compression....maybe I can get some measurements on mine and we can compare...
thanks for the thread though...info is good...I'll just watch this and see if I can apply anything here to mine...If you think I could be of help ...let me know.

later
ts

open your unit up and apply a resistor in parallel on top of R2 until you get the meter to sit at 0dB in GR meter mode. I dont even know what value I wound up with, but it works and doesnt seem to effect the circuit otherwise with it in or out. Also noticing on both of mine that the meter in output mode has 4dB insertion loss, guess you have to use it in GR mode. I would also reccomend putting a threshold pot on the front panel and moving the wiring, the way this box works, the threshold and input attenuator are fairly interactive, keeping the threshold on the back really cant be practical.

dave

C1, or something on that node, is shorted to ground. Or R11 is open.

here's my copt of the schematic if this helps



later
ts

OK this may or may not have anything to do with your problem but...
i see here from your posts that you have replaced so much in this unit...for the better...(you still have the cool looking meters ...right...ok just kidding)
and I am reminded of a RCA unit that I had.
the power switch turned out to be the problem... I was getting some weird measurements...I forget the numbers but...whatever the measurement on one side of the switch it ....should be the same on the other side when turned on....well it wasn't.
a friend explained it to me something like this...
the metal on the switch had oxidized and the switch was now acting like a diode. bypassing the power switch solved the problem.
I offer this because it seem to be one of the only parts you haven't replaced and I had this weird experience...maybe it's related maybe not.
hope this helps
ts

I've been looking at this schematic some more and...
as it relates to insertion loss...
do you think that this was planned/designed into the unit so that if the unit was bypassed during broadcast before a transmitter it would automatically give you some headroom so the transmitter would not blow.
assuming this:

could changin r12 remedy the insertion loss?
just a guess....

thanks
ts

ok, Ive made some headway, sort of.

Under the turret board is a gigantic mica 1uf cap in a big molded metal box, C1. I removed this and replaced it and all the hiss went away. Ive got some decent compression happening now as well. The hum however, is still identical but the unit is very quiet with the threshold all the way up. When the threshold lowers, the hum comes up. Oddly enough, I had an altec 438 (also a mu limiter) years ago that had the exact same problem, quiet when not compressing, but lots of hum when it went into compression. Does this point towards a bias problem maybe on V4? The manual shows 140V on pin 6 of V4, I measure 183V, does 23 volts seem like it might have these kind of artifacts(hum)? How would I go about knocking that down to test my theory?

what is the function of C5 in this circuit?

The metering is now even more F'ed up since I changed C1.

there are two additional components in both of my federals that arent on the schematic, I dont have a program to draw it, so bear with me-

The attenuator side of C1 at the junction of the attenuator side of R1 has a 100K resistor in series with R11. At the junction of the 100K resistor and R11 is a .47uf/200v cap to ground. I was tempted to remove this, but my unit that is working well has the exact same thing, so left it. Any ideas on what those two parts are doing?

PRR- I checked around C1 and cant find any shorts. R11 measures 47K in circuit.

Toobie- I will check the switches next time I work on this. Try a 100ohm resistor in parallel with R2, I think thats what I have in there, they measure 40 ohms together in circuit, that zeroes the meter in GR mode on this one. With or without that resistor, the meter appears to track the same so far as I can judge testing by talking into a mic.

dave

Did you check your power supply for ripple yet? Any ripple will be amplified by V4 and injected to the grids of V1 and V2. It will be cancelled out to some extent, being common-mode, but the feedthrough could get worse as you turn down the threshold control unless V1 and V2 track perfectly throughout the range of gain reduction, which they won't.

what is the function of C5 in this circuit?

Click to expand...

It's a coupling cap. It keeps the plate voltage of V2 off the threshold pot and the grid of V4.

Any ideas on what those two parts are doing?

Click to expand...

If I'm picturing it correctly, based on your description, it's a mod to slow down the attack.

While you're digging around that part of the circuit, lift one end of R1 out of circuit and check R1 with an ohmmeter. If that resistor had gone high or opened, the grids would be effectively floating at DC and very low frequencies, which would cause or exacerbate hum problems as well as throw off the grid bias on V1 and V2.

> Is it normal for a meter to cause a -4dB insertion loss??

No. You may have a shorted diode in the meter bridge.

However, you should be running this with a 600 ohm termination. Running it unterminated is going to give unintended results.

> in order to get the GR meter to sit at 0 (instead of off scale) I needed to add resistance in parallel at r2.

BAD IDEA. Yes, trim R2 so the meter don't melt, but put it back the way it was when you fix the problem that was causing excess V1 V2 current.

Obviously your C1 was bad, shorting-out the diode-bias that holds V1 V2 grids about a volt negative. You may also have a short in the added 0.47uFd, or opens in the joints or wiring in R11, R11A.

> {R2} in or out it didnt affect the output level or sound of the unit, only where the needle sat while in GR metering mode.

Of course not. Why would it? It measures the DC current. Ideally there will be zero audio current/voltage in R2; push-pull cancellation. If R2 were huge, over 10K, it would have some little effect on the GR curve, especially the knee. But it is only 240 ohms. Its drop from max-gain to max-GR is less than 1.2V, which is negligible here.

While it is running, no-signal, put on your leather gloves and yank V1 and V2, one at a time. If the GR meter shows 100%/0VU with both tubes, it should show 50%/-6VU with just one tube, either one (after warming). Or to say it another way: you should be getting 5mA with both tubes, 2.5mA with just V1, 2.5mA with just V2. Maybe you are not getting 5mA, but still either tube alone must be half the total. A 10% difference is fine, 20% OK, 30% or more means the tubes are mismatched enough to cause big problems (distortion in deep GR, but also crappy buzz rejection).

And pay attention to what Dave is asking:

I still strongly recommend taking the defective cap can out of the circuit.

Did you check your power supply for ripple yet?

Because of push-pull action, power buzz will do crazy things.

thanks guys for the continued comments.

the next time I sit down, I will entirely remove the old cap can. I have a real embarassing question, but how do I check for ripple? I have a scope here.

I will also check R1. Going through the manual, it appears that C1 and R1 are fairly critical to the operation of the circuit, knowing what I know now I would have gone to those two parts directly.

Looking at this further, it looks like the threshold for the unit is actually the control labeled "current control" and the pot labeled "threshold control" is more like the ratio and the value of the pot is IMO too high, after about 2/3 rotation the unit runs out of makeup gain.

CR1 isnt made up of regular looking diodes, its an odd looking thing which Ive never seen before, looks like four wires coming out of a terminal block. I want to get rid of it, which type of diodes would be appropriate to use to rectify the meter? Ive never done that before.

Adding series resistence between R1 and R11 in theory should slow the attack time?

thanks for all the help on this, its really appreciated.

dave

> how do I check for ripple? I have a scope here.

Read the voltage ratings on your scope and probe!!!! I built my own probe when I was working on high voltage amps for a living.

Though in this case, you have a shortcut. The top of the 1K pot is a fraction of the power rail, and has no additional filtering. So read the AC voltage there, and multiply by the resistor ratio or by the ratio of the two DC voltages, rail and top-of-pot.

> it looks like the threshold for the unit is actually the control labeled "current control" and the pot labeled "threshold control" is more like the ratio

It is a crude limiter. It does not have a well-defined ratio or threshold. And I dunno what a "current control" is supposed to do.

However the "current control" is your main control of knee sharpness and ratio. If you set it to zero volts, it has a very soft knee and low ratio. For limiting, you have to set it to a fairly high voltage. For high-ratio limiting, maybe near 20V. It wasn't made to do that, as shown by the pretty lame control amp. I suspect 2 or 3 volts gives a good range of no-action, and above that a range of lower dynamic range. If you are compressing clean tracks for effect, maybe you want compression to start early and soft, so 1V might be your setting.

The "threshold" control is OUTPUT level, not input level as many modern boxes do it. And it isn't user-adjusted. The technician sets that so the output level while doing 6-10dB of GR is just tapping zero VU, then tweaks a few dB until he hears the transmitter distort. In your application you would just trim for zero VU output. Then the amount of "smoosh" is set by the input level pots and by the output level from the control room console. With the input pot all the way up, input threshold should be well under 0.1V, so you should have no problem smacking all the GR it is capable of.

Don't know if this helps, but how I have interpreted the manual and used my two units was that "Threshold" on the rear was more like a ratio setting while the "Current Control" was actually the threshold setting. The knob on the front is input attenuation at which I never really go beyond 6.

I love the barrier strip hookup on the back as it's very versitile with regards to going balanced or unbalanced using "CT" or center tap.

This unit is amazing on Bass.

I send signal to it via a channel fader and return to another channel on my console.

How easy would it be to add and output control, is it just adding a pot?

You'd use one of these.

They're kind of hard to find, but they do turn up now and then.

There's also a DIY option using common switches:
http://www.groupdiy.com/index.php?topic=2229