Print

[2E0ILY]

I have a used Alpha 86 HF amateur band amplifier on the bench. It appeared to be working fine when I received it, but subsequently I have found an issue that may be related to the PIN diode RF switching that is used to enable fast CW with break in.

It has two operating conditions. STANDBY where RF goes in to the antenna socket and out to the transceiver via the relay K3. So when this relay is closed RF comes in from the antenna, via the closed relay contacts, to the transceiver. in this state RX is fine. In its OPERATE mode K3 opens and RF is sent via, as far as I can read the schematic, the RX PIN diodes D3, D4, and D5. In this state the RX signal is attenuated some 3dB. More so on the higher bands.

My first question is, does PIN diode switching in such amplifiers cause attenuation of the receive signal, normally?

But I have just found another anomaly, in the OPERATE status if I move the TUNE and LOAD capacitor positions from where the amp is happy to load up and TX into a dummy load, but not transmitting, just receiving via the transceiver, the attenuation reduces or virtually goes altogether, depending on band. As such I am wondering if the big, rare and expensive UR4001 stud mount PIN diodes are leaking the RX signal back in to the tank circuit? I could REALLY do with someone that understands the schematic and PIN diodes far better than I to advise. If you didn't hear the drop in RX receive sensitivity, or look at the S meter, you might think, as I initially did, there was nowt wrong with it.

I find it hard to believe the PIN diode set up normally attenuates 3dB!

I was tempted to convert it to two vacuum relays and bypass the fast PIN diode switching, a fairly common mod, but I would rather leave it in the more valuable original state if possible. I attach two schematics, if anyone would be able to find the time and inclination to study them I would be most grateful as I don't want to spend a lot of dosh on new stud mount PIN diodes if it might be just D3, D4 and D5. These are also near unobtanium, but Alpha now say it's OK to replace them with "low capacitance" 1N4007's. I can find nothing lower than 8pF ones, so i guess they mean those. Alpha have shut for the virus with no signs of opening again any time soon.
Thanks a lot!

The two circuit diagrams I feel are relevant can be found at

http://www.chriswilson.tv/alpha/diag1.jpg

http://www.chriswilson.tv/alpha/diag2.jpg

[W9IQ]

Have you confirmed the proper reverse bias of D1 and D2 during receive? This would involve checking for the 28 volts at the output of L1 (schematically the top end) and that R1 and R2 are the correct resistance. Also check C1 and C2 for shorts or leakage - these caps are high stress items. If you don't have a way of properly testing these caps, you may consider simply replacing them as they are low cost. Also check that NOT(R) is within 0.6 volts of ground when in receive - but be careful because this rises to 750 volts during transmit (or a failure in the logic circuit!).

- Glenn W9IQ

[2E0ILY]

Have you confirmed the proper reverse bias of D1 and D2 during receive? This would involve checking for the 28 volts at the output of L1 (schematically the top end) and that R1 and R2 are the correct resistance. Also check C1 and C2 for shorts or leakage - these caps are high stress items. If you don't have a way of properly testing these caps, you may consider simply replacing them as they are low cost. Also check that NOT(R) is within 0.6 volts of ground when in receive - but be careful because this rises to 750 volts during transmit (or a failure in the logic circuit!).

- Glenn W9IQ

No, due to the cost of the stud mount diodes and their rarity I have held back replacing anything until such a time as I hopefully received some sound concrete methods to test the circuit correctly, Thanks to your excellent reply, (for which I am extremely grateful), I have a way forward to follow with some confidence. I did some reading up on PIN diodes last night and in my notes I put check the bias, and you now show me just how to do that, thanks a lot!! I now have a pair of UM4010DR stud PIN diodes, but would rather try and ascertain the originals are actually faulty before removing the originals as they look tricky to get at and remove without damaging anything else. I have had bad experiences with elderly PCB tracks...

Would ceramic caps be appropriate for C1 and C2, or should they be something else, perhaps silver mica? If ceramic do they need to be COG type? Thanks again Glenn, I have a bit more hope to fix this now without engineering a vac relay set up.

[W9IQ]

Silver mica would be fine. The main objective is very low ESR (effective series resistance) as these capacitors pass relatively high amperage (~5 amps each). You could also use ceramic C0G capacitors. As a class 1 device, they generally have low ESR. But make sure to check out the SRF (self resonant frequency) as it may prove unusable over the full frequency range of your amplifier.

Good luck with the testing. Keep in touch with your progress.

- Glenn W9IQ

[2E0ILY]

Silver mica would be fine. The main objective is very low ESR (effective series resistance) as these capacitors pass relatively high amperage (~5 amps each). You could also use ceramic C0G capacitors. As a class 1 device, they generally have low ESR. But make sure to check out the SRF (self resonant frequency) as it may prove unusable over the full frequency range of your amplifier.

Good luck with the testing. Keep in touch with your progress.

- Glenn W9IQ

Thanks again Glenn

I couldn't find any SM caps of that value that wer not huge and hugely expensive so have ordered some COG MLC caps from Mouser with very low ESR. Their resonant frequencies are not stated in the spec sheet however. I will replace R1 and R2 at the same time with metal film resistors, 1% to hopefully balance D1 and D2 bias.

I have this question however, brought about by a reply on another forum:

OK, so I have an Alpha 86 amp on the bench, an amp that uses big stud mount PIN diodes for TX switching. In STANDBY mode the path from the antenna to the transceiver is via an open frame relay. In OPERATE mode it is via PIN diode network and the RX signal is attenuated some 3dB. On another forum two people who use PIN diode switching on a commercial level for RF comms have stated such loss is not unusual and actually quite good... They say amp manufacturers plug the power output but never mention what PIN diodes do in the receive path. So, anyone with an HF amp that uses PIN diodes in the receive path, can you hear and measure on the S meter any attenuation when the diodes are in the receive path, and have you seen this mentioned in the pros and cons of such a switching method?

Thanks

[W9IQ]

Hi Chris,

Since the capacitor SRF was not published, make certain to watch for anomalies in the upper HF bands once you have everything working again.

The 3 dB receive loss would be high in my opinion but I have not tested a good version of that amp to say for certain. The telltale issue in your case is that the output network is affecting the receive signal. This suggests that D1 and D2 are not providing sufficient blocking during receive. That is why you first want to check all of the bias points related to those diodes.

Once the D1/D2 diode function side is checked, you would then move onto the receive side diodes. In order for them to work properly, they need to be forward biased with sufficient DC current to make them operate at a low loss. The 5 volt supply through L2 and L3 provides the bias voltage for D3, D4 and D5. R3 and R4 are the current limiting devices for this forward bias so they are being forward biased with about 10-15 mA of current. They are reverse biased during transmit by about 750 volts from the NOT(R) signal.

The function of D6 is to further shunt the receive path during transmit. It becomes forward biased during transmit. It is reverse biased  from NOT(T) during receive to ensure low loss from this path.

Finally don't overlook the relay as a source of receive loss. Dirty contacts could easily add 3 dB of receive loss.

I know that is a lot of information but take it one step at a time. I am happy to stay with you throughout the process. Just remember that there is relatively high (lethal) voltage present in these circuits even when not transmitting. And the high voltage circuit is live also. So exercise extreme caution and use properly rated test equipment and cables. Even if the amp is powered off and unplugged, the cap bank can still hold lethal voltages for several minutes (and more if there is a bleeder/balancer chain failure). So always discharge with a "chicken stick" before assuming the un-powered amp is safe.

- Glenn W9IQ

[2E0ILY]

You are a star Glenn, thanks a bunch Sir!

Can I insert a decent multi meter in the path between L1 and R1 /R2 to measure the bias current? And next between L5 and R3 / R4? Should I also change C7 which is a parallel pair of 0.1uF caps? I bought 4 of them in case.
 
I don't see the relay K3 affecting it.... Ooops, yes, of course I do, the other contacts... I did clean these to no effect. But I'll check those again.

I ran the amp last night and I have to say I have a *FEELING*   things were better on RX, so maybe, just *maybe* the contacts are high resistance in OPERATE and they are cleaning themselves up with use. I think this amp has been unused for quite some time.

[W9IQ]

Can I insert a decent multi meter in the path between L1 and R1 /R2 to measure the bias current?

There will be very little to no current since these are reverse biased during receive. It should be sufficient to measure the voltage (to ground) from the top of R1 and R2. You should see a little less than 28 volts in both cases. If you measure any voltage each resistor, then you probably have a bad D1/D2 as this indicates current is flowing through a reverse biased diode.

And next between L5 and R3 / R4?

In this case, I would simply measure the voltage across R3 and R4 during receive. Since they are 33 ohms and there should be 10-15 mA flowing through them, you should see a voltage of about 0.3 to 0.5 volts across each of them.

Should I also change C7 which is a parallel pair of 0.1uF caps? I bought 4 of them in case.

There would be no problem doing that although if you do a lot of changes, you may later start questioning if your changes introduced new problems. Their purpose is to keep the 5 volt supply isolated from the antenna/load. You could test if they are shorted by disconnecting the antenna/load and simply measuring from the center pin of the output connector to ground during receive to see if there is any voltage there. If this is the case, that could affect your receive signal strength as the diodes may not be sufficiently forward biased to assure low loss in the receive chain.

It would be nice if it were just the relay contacts. Be gentle when cleaning so you don't disturb the plating...

- Glenn W9IQ

[2E0ILY]

OK the C7 test is dead easy as the cover is back on. In OPERATE I see 41mV DC on the antenna socket, in STANDBY zero volts. Is 41mV enough to concern myself with please Glenn?

[W9IQ]

I wouldn't think so, Chris.

- Glenn W9IQ

[2E0ILY]

OK, thank you, it was as easy a test as they come so no worries

[W1VT]

Given the difficulty of working  on an amplifier and the desire to keep it as pristine as possible, how about building a duplicate circuit and experimenting on that instead?
You may want to start of with the basic PIN diode switch and measure its loss.  Then add additional circuitry.

Zak W1VT

[W9IQ]

Zak,

You should obtain about 0.1 dB loss per diode when using 1N4007 diodes for the receive chain. The Alpha uses 3 diodes in the series receive chain so I would expect a loss of less than 0.5 dB. You would also think this would be well known in the industry if 3 dB was normal since simply switching the amp to standby would make any high receive loss through the PIN diodes very obvious.

- Glenn W9IQ