833a modulator distortion

[KB5MD]

I am in the process of reconstructing a BC-1j Gates into a 19" cabinet.
Everything is completed and operating, but for the life-of-me, I cannot get the distortion out of the modulator.  The audio is clean and clear up to the modulator and then turns to crap.  The modulator is a (2)x833a p-p running 2800 volts on the plates and grid biased at -150v just as it was in the original Gates.  I am open to any suggestions as to what is going on or rather, not going on.

[k7yoo]

I am not a Gates expert but wonder if all is well with the mod transformer connection to the driver section. Both the driver and the PA were modulated on the BC1 that I had. The driver had gone up in smoke at one time and the PCB was pretty ugly.
Skip

[W2PFY]

At that voltage 2800 I would sugest you adjust your bias voltage on each tube so it will draw about 75 mills per tube. With -150 volts on them. they are biased at cutoff.

[W2XR]

At that voltage 2800 I would sugest you adjust your bias voltage on each tube so it will draw about 75 mills per tube. With -150 volts on them. they are biased at cutoff.

That's right.

In my homebrew 833A moduiator running 2600 VDC on the plates, the grid bias is about -60 volts for an idling current of 50 Ma per tube. You are biasing the tubes into cut-off and the very non-linear region of operation. Get that bias right and the rig will sing like a bird!


73,

Bruce

[KB5MD]

Well, I did as Bruce suggested, cut the bias voltage to -60, adjusted
current to about 50Mils per tube and the distortion just will not go away.  Where can I find a schematic of the homebrew 833a modulator? If I only had something to compare to, I think I must be missing something. Sort-of "the forest for the trees type of thing.

[W2XR]

It could be several things at this point.

Are you sure the audio driver stage and the driver xfmr are good? You can simply test this by connecting an amply rated 115 VAC to 6.3 VAC filament xfmr to the output of the driver xfmr and listening to the audio quality thru an 8 ohm loudspeaker. The 115 VAC side goes to the secondary of the driver xfmr. If the audio is clean at this point, the problem is obviously within the modulator stage itself.

Make sure that the DC blocking capacitors that couple the negative feedback from the 833A plates to an earlier voltage amplifier stage within the audio driver are not leaky or shorted. There are two of these voltage divider/feedback ladder networks that provide the correct amount of out-of-phase audio voltage to the driver for feedback. Any DC present after the DC blocking capacitor will severely mis-bias that stage. You can only check for this when the HV is applied to the 833A plates, but why do this? Just replace the caps! If you have'nt done so already, it's good practice to replace all of the caps within the driver as a matter of course.

Another possibility is a bad modulation xfmr. If you get large excursions in modulator plate current, but low percentages of modulation, usually the mod xfmr is suspect.

This should be a fairly simple thing to troubleshoot. Just watch the 2800 VDC when doing this testing!

Good luck.

73,

Bruce

[k4kyv]

I ran into an interesting circuit in a schematic of an RCA 5 and 10 kw transmitter.  The Gates BC1-T and later transmitters used a tertiary winding on the modulation transformer to apply some modulation to the rf driver stage.

The RCA circuit used a tap on the modulation transformer secondary, near the bottom end.  The bottom end of the secondary went to the MV power supply that powered the driver stage.  The top end went through the audio coupling capacitor to the top end of the modulation reactor.  The tap went to the plate of the driver stage.

The mod tranformer  secondary carried the DC to the driver stage, but the final amplifier plate current was shunt fed through a modulation reactor.  The ampere-turns in the portion of the secondary that carried the DC to the final was insignificant enough that it didn't saturate the core of the mod transformer.

It should be a simple matter to use this circuit in a ham transmitter with a multi-match modulation transformer similar to the UTC (C)VM-5.  Use one of the normally unused taps near the bottom end of the secondary winding to feed the driver stage.

[flintstone mop]

I'm trying to think of the plate modulated scheme and we haven't mentioned the feedback ladder. I know if it's not connected correctly, there will be howling type feedback. If it's the troublesome 807 audio board, look for out of tolerance or burned up resistors. 807's in good shape? Did your distortion start AFTER re-assmbly in the 19 inch rack??

Fred

[W2XR]

I'm trying to think of the plate modulated scheme and we haven't mentioned the feedback ladder. I know if it's not connected correctly, there will be howling type feedback. If it's the troublesome 807 audio board, look for out of tolerance or burned up resistors. 807's in good shape? Did your distortion start AFTER re-assmbly in the 19 inch rack??

Fred

Fred,

The BC-1J that this gentleman is troubleshooting does not use the 807 cathode follower arrangement (as used in the Gates BC-1T, BC-1G, & BC-1H rigs) that I believe you may be thinking of. The BC-1J uses a quad of push-pull/paralell 1622/5881s with fixed bias into a discrete pair (yes, a pair) of driver transformers. Each driver xfmr handles one push-pull phase and drives the grid of it's respective 833A modulator tube. Definately one of Gate's more kludgey and complicated designs IMHO, although it must have been significantly less costly and demanding of real estate when compared to the beautiful sounding and really robust push-pull class A 845 audio driver as used in the magnificent Gates BC-1F. I use the BC-1F 845 audio driver circuit to swing the grids in my homebrew push-pull class B 833A modulator; a nearly identical circuit is used in the classic Raytheon RA-1000 rig as well.

Later, with the discontinuation of the BC-250K, BC-500K, and BC-1J, Gates dropped this quad of 5881s mess and went with the cathode follower configuration, and in so doing, they were able to dispense with the problematic and costly pair of class B driver transformers and separate modulator bias supply.

73,

Bruce