I'm restoring an Emerson FU-427 portable and have re-capped and replaced all out of tolerance resistors. I did the main two electrolytics first, and powered on to see how healthy the radio was. I got basically oscillation and howl. I re-capped the paper caps starting from the RF section back to see if I could catch the bad part. No luck.

- The 1N5's, 1A7 and 1H5 all have shields.
- All tubes have been tested and also swapped out.
- When I touch a screwdriver to the top of the 1A7, the howling stops and I can tune up and down the dial.
- When I remove the screwdriver, the station remains tuned and there is no oscillation or howling, but noise almost sounding like a microphonic tube is heard
- If I then try to tune another station, the howling/osc starts immediately
- I checked and re-checked the wiring and tested poking/moving the wiring in case it was a position sensitive wire.

Schematic: http://www.nostalgiaair.org/resources/110/M0005110.htm

I found a service Note: To correct a bad howl when the volume control is turned down, replace the 40mfd. 25-volt condenser which is mounted below the main filter. Intermittent operation may be caused by a defective main filter (20-20-150).

The original(?) 40uf cap (C30) was connected Positive to chassis ground. The schematic does not indicate polarity of the electrolytics. Does this look correct? Any other thoughts on what to look for on the howling?

Hi

Positive should go toward filaments. Negative to chassis/ground. They might have got away with this cap being reversed since it only has low voltage?

Your RF and IF stages have high gain and are near oscillation. Be sure bypass caps are mounted near their original locations. Are tube shields making good contact with the chassis? You may need to slightly detune an IF Transformer to stop oscillation?

Thanks Norm - I will try a fresh cap the other way. Shields are tight and the bypass caps are nearly spot-on to original locations. I've also pushed them around a little while running with no change. I did try tweaking the IF slugs and oscillator trimmer a little and it did improve the unwanted oscillation, but also the sensitivity (as expected).

I'll throw the sig-gen on there also and see where it lands. The IF on this one is 262 khz.

Looks like the 40 uf cap was not original, but a replacement with correct polarity did not fix the issue. I'll keep digging.

A high value resistor across one of the IF coils might drop the Q a bit to get out of the oscillation zone.

Thanks - I'll consider that if I don't find the issue. I found this set at an antique store, and was going to pass - but aside from the style that attracted me, when I peeked inside it has a very large speaker for a portable. A six inch I believe - and the tone as expected is very nice. So - I do want to get this one working well!

I should also clarify that mine is actually an FU-428 (same chassis, different case).

Mine looks like this but actually much cleaner:

So it's been awhile... finally pulled this back off the shelf. I still have the oscillation, and high B+ (114v at the plate of the 1A7, that should be 80v). All resistor values have been checked, caps replaced, etc. My testing is done with AC input at 117v. Voltage at the 117p7 rectifier cathode is 130v, which is correct per schematic. R4, R5, R7 have been checked, replaced anyway. Tubes are all NOS and have been swapped with others.

One separate, but maybe related question: C30 (highlighted yellow below) and referenced in thread above - As it came, positive of this 40uf, 25v electrolytic was tied to ground. It did appear to have been replaced, so probably was a mistake - but can someone take me through this circuit to validate?

Some more things to try:

Is there a bypass cap at the junction of R1 and R12 in the 1N5 AVC line? If not, try adding one.

You might take a 0.01 to 0.1uF cap with long leads, and bridge between various points and ground. Try points in the AVC circuit, B+ distribution, and the filament string. Be careful; a slip can pop a tube filament.

Try measuring the resistance between the RF and converter tube grid caps. Do you get the expected 8 Megohm value?

Measure between the IF tube grid cap and ground. Is this 5 Megohms?

With power on, measure the voltage across R3 ( below T3 ). Use a very high impedance meter. Try to get values when the howl is going and when it is stopped by the screwdriver trick.

See if the B+ at the 1A7 stage changes when you make the howl stop with the screwdriver.

Look around the IF stage again. The symptoms seem to point to this stage. This stage has no AVC; if it oscillates the AVC on the RF and converter stages will kill the gain there, and reduce the plate current for those tubes.

Try a capacitor to ground from the converter grid cap. See if this prevents the howl. It should kill any radio reception.

Let us know what you find.

Yes. C30 should have its negative end to ground.

Ted

Thank you for the detail Ted! I will work on this tonight and report back.

I was able to try a few tests tonight. See below.

1. Is there a bypass cap at the junction of R1 and R12 in the 1N5 AVC line? If not, try adding one.
There is not a bypass cap here. I tried a .02 at the junction of R1 and R12 to ground and there was no change.

2. You might take a 0.01 to 0.1uF cap with long leads, and bridge between various points and ground. Try points in the AVC circuit, B+ distribution, and the filament string. Be careful; a slip can pop a tube filament.
Didn't try this tonight.

3. Try measuring the resistance between the RF and converter tube grid caps. Do you get the expected 8 Megohm value?
8.6 Meg

4. Measure between the IF tube grid cap and ground. Is this 5 Megohms?
5.1 Meg

5. With power on, measure the voltage across R3 ( below T3 ). Use a very high impedance meter. Try to get values when the howl is going and when it is stopped by the screwdriver trick.
Oscillating, I got between -1.5 and -2v dc, positive of my meter to T2 and negative to the junction with C18. No oscillation (tuned to a strong station) showed 0 volts.

6. See if the B+ at the 1A7 stage changes when you make the howl stop with the screwdriver.
B+ stays at ~105v whether oscillating or not.

7. Look around the IF stage again. The symptoms seem to point to this stage. This stage has no AVC; if it oscillates the AVC on the RF and converter stages will kill the gain there, and reduce the plate current for those tubes.


8. Try a capacitor to ground from the converter grid cap. See if this prevents the howl. It should kill any radio reception.
This does stop the oscillation, and reception.

I also installed a new 40uf cap correctly at C30. Out of curiosity, I also measured the filament string (ground to the junction with C30) and measured 5vdc.

I think the high B+ you are seeing is due to using AC line power while making the measurements. The Rider schematic lists voltages taken with a 90VDC B+ source. All of these voltages will be higher when the radio is run off AC line power.

The A voltage across C30 appears to a bit low, at 5.0V. 5.6V would be the nominal value. Increasing this would not help the howl situation, though.

The Rider voltages are also specified as being measured with a 1000 ohms per volt meter. This will load down most of the points measured and produce a lower reading than you will get with most modern meters. So there is every chance that the B+ values you are seeing are perfectly OK.

Since a capacitor from the mixer grid to ground stopped the oscillation, it looks as if this tube is involved in producing the howl. Perhaps it is not due to the IF stage, after all. You might try the same thing with the capacitor, to RF tube's grid cap.

It's still not apparent to me what is causing your howl. The results of your tests, so far, seem to indicate that the RF and or the converter stages are involved; that it is not just the IF stage "taking off" by itself.

Perhaps you can find the undesirable feedback path when you probe-around with the extra bypass capacitor. Be sure you try it across C30. Electrolytic caps make poor RF bypasses, and the manufacturer's notes indicate that there was some sort of problem back in the day.

One more thing to check would be for continuity through the loop antenna. Be sure that the left hand end of C11 on the schematic reads a low resistance to ground. Without the tuned circuit working, there would be more gain at the IF frequency, and that might cause trouble.

Has anyone been in this radio since it last worked? Could there possibly be alterations that did not turn out so well? There are several things you might try to make it stop howling, but most would amount to modifications that should be avoided, if possible. It would be best to actually locate the fault, rather than to patch-around and possibly make things worse.

Keep reporting back. Someone will have more ideas.

Well, good luck. Should be a nice radio when you get it working.

Ted

Thank you for all the thought and ideas. I will try lowering the AC voltage to 90v to see what that does. I've been using a Fluke 87v, but will also dig out my old VTVM also and do some comparisons.

I will keep working it and report back.

Mark

I noticed that the screen grid of the 1N5 IF amplifier is not bypassed. Maybe try a .05 uF capacitor from the screen grid (pin 4) to ground. Also check that the .05 uF screen bypass for the 1N5 RF amplifier is located close to the tube. On the schematic it is very far away, but of course that is no clue to its physical location.

-David

Here is just a thought based upon what you said about the oscillation stopping when you touch a screwdriver to the 1A7 grid cap.

Touching this point adds a tiny amount of stray capacitance to ground. Really tiny. This somewhat detunes the input RF transformer.

Have you considered a minor tweak of C4 to see if this might also stop the oscillation (also a tiny change of capacitance to ground at the grid of the 1A7)?

Curtis Eickerman

I've somewhat ruled out tweaking the IF alignment (slightly out of peak) as a fix, as when I tried, it took a lot of change to get the oscillation to subside - rendering the radio nearly unusable. So, that pointed me to look for a serious functional issue rather than maybe a design anomaly. Good thought though!

Well, it was worth a try. It still sounds like the 1A7 is self oscillating other than just as it should as a Local Oscillator.

Given that C12, C14, C15, C17 and C28 are all in place and OK, another option is increasing the value of R5 (1K) plate resistor and/or R4 (30K) screen resistor to reduce the gain in the 1A7 without crippling it (you already noted the voltages are on the high side anyway).

You could probably go as high as 2X on either or both. Going up certainly won't hurt the tube and might give you some hint as to what is going on.

Curtis Eickerman

Good - I will try that also.

I am curious to try lowering the AC voltage to 90 to simulate battery voltages and see if that makes a difference. I don't have it in front of me, but thought the Rider document said the voltage table provided was based on 117v. My other concern with the high voltage, if it turns out to be normal - is that the 1A7 for example, the maximum plate voltage listed is 110v (from memory) and with the input voltage set at 117v, I am uncomfortably close at 105v. I will drag out my VTVM tonight to recheck this.

Don't forget that lowering the AC line voltage will also drop the filament supply, "A" voltage.

You could temporarily connect a 6V battery across C30 to bring it back up while testing. Four flashlight batteries in series would do.

In any case, I'm not sure what this exercise will accomplish. I think the higher B+ is normal on AC line operation. Measure at the rectifier, compare this with the 90V you would get with a B battery, and increase your expectations for all the other B+ values by about the same amount.

If you reduce B+ enough, the howl will stop, but the radio will not play so well, if it works at all.

If you want to lower the gain of any of the RF or IF stages, some negative grid bias will work well, since they are all remote cutoff types.

Ted

OK - you guys have convinced me on the voltage. I did find something that may be telling. If I put my finger on the grid of the 1N5 rf amp tube, it quiets right down and the radio tunes very strong all over the dial with no oscillation. I just went through all the antenna wiring again and all looks good per schematic. R1 and C11 are on the antenna itself. I also checked the antenna loop resistance and see 1.5 ohms.

Quick update - if I put a 68pf cap from grid to ground, same effect as above.

Ted

I think that you have all of us scratching our hair. Your troubleshooting keeps eliminating things that might be the cause of your howl. It does sound as if the RF & converter stages are where the howl is being generated.

You might get the howl going, then see if you can tune the resulting RF emissions with a second radio. It would be interesting to see if the RF frequency tunes up and down as the Emerson is tuned.

If you do detect the howl on the other radio, try a bypass cap on the plate of the 1A7 converter, then the grid cap of the IF tube. See if the capacitor in either location kills the RF picked up by the second radio.

Have you had a chance to try an additional bypass cap on all the non-signal points around the RF and converter stages?

When the howl is going, how much AVC voltage do you see? ( Gotta use a high impedance meter for this measurement.) The clockwise end of the volume control is the source for the AVC voltage. It should be several volts negative. The grids of the RF and IF tubes should be even more negative, due to the very large grid leak resistor values. Try measuring across R11 and R12 with a VTVM or high impedance digital meter. The volume control end of each resistor should be positive relative to the grid end.

Have you tried different tubes? At the least, you might swap the RF and IF tubes. The grid bias for the RF tube is very dependent on the grid current, and that is often not very consistent among different examples of the same tube type. And the DC bias voltage controls the RF gain.

If you end up having to knock-back the gain a little, another way would be a resistor across the RF tube's filament. This would not hurt the selectivity or tracking of the front end, while helping to bring the voltage across the other tubes up to the specified nominal value.

Ted