HELP - Larcan 6 Meter Amp misbehaving

[eric gruff]

Fellow Larcanites (Larcanians?),

I purchased a LO/LO amp from VO1KS, and in the process of setting it up may have damaged something. VO1KS and I have spent a lot of time trying to figure this out, but we're pretty much done with any ideas, and before I turn a fairly expensive investment into a footrest, I am hoping one of you can help.

There is a long story involved, but I will stick to the basics of where I am right now.

1. FETs seem fine. I have 500 mA bias on each on at idle and the 7 Amp fuses on the amp are intact and do not blow during any of my attempts to use the amp.

2. Antenna is a SteppIR that is tuned to 1.06:1 SWR on 50 MHz (I use both CW at 50.090 and JT65 at 50.275 MHz). I have three SWR meters, including a Bird 43 that are consistent with all power and SWR readings.All output coax is new LMR400 that routinely handles 1500 W on 160-10 M from my Alpha 9500.

3. Two separate HP 3 kW power supply units (PSUs) have been tried. I just bought a brand-new, never used one and it behaves identically to the first one. These are the ones that all of you are using - http://www.k8gp.net/station-equipment/hp-3000w-power-supply-info. Wired as in the picture on K8GP's Grid Pirates page from a very steady 240 V wall supply (powers my 1500+ W Alpha 9500 without issue).

4. DC and RF lines have toroids (good quality) and snap-on ferrites, respectively, on all of them. The DC B+ and B- each consist of a twisted pair of 12-Gauge wires, all run around a ferrite.

5. I have checked all diodes, resistors and capacitors, including lifting all power resistors from the board to be sure I didn't smoke them.

6. Power into the Larcan is carefully monitored (ever since the first time I tried it, which might have been the issue that one time I didn't), as are voltages.

Here's my test result summary. Very reproducible and very frustrating:

- Power up amp. Total DC draw about 3.8 Amps (8 x 500 mA, give or take), B+ voltage 51.2 V, bias 38.9 V. (so far, so good)

- Add 1 W drive via 6 dB external attenuator: Power out about 100 W, current draw about 12 A, B+ voltage 53.3 V, bias 38.9 V (note increase in B+)

- Add 2 W drive via 6 dB external attenuator: Power out about 200 W, current draw about 18 A, B+ voltage 56-57 V, bias 38.9 V (note additional increase in B+)

- Add 2.5 W drive via 6 dB external attenuator: Power out just under 300 W, current draw about 28 A, B+ voltage 62-63 V, bias 38.9 V 

- Increase to 3 W drive or higher - PSU shuts off B+ and green light on the front of PSU starts blinking

- Rinse and repeat

Before I started using a clamp-on ammeter, I had a shunt ammeter from eBay (the exact one that AC0C uses - http://www.ac0c.com/main/page_acc_shack_larcan_kw_6m_amp.html). It was in the negative DC line and was reading insanely high current readings (up to 72 A) when I pushed the drive to 3 Watts and got 300 W output just before the PSU said "no mas" and turned off. It was not showing a voltage change, but seemed to show a massive power draw from the amp.

VO1KS thought the ammeter was the issue, and I stupidly tried to use it in the B+ DC line, so now it is a crispy critter. However, if the PSU was stopping above 3 kW DC out, where was all that power going? Now, it seems like the B+ is being pulled very high with the same result, but NONE of the 7 A fuses has ever blown, and when I put my DVM across each fuse as an ammeter, they only get up to 3 Amps or so each. So, the power is bypassing the FETs and going somewhere, but WHERE???

I will be eternally grateful to anyone who can shed light on the subject before my frustration level goes any higher and I take the Larcan out to the backyard for target practice. I actually went out and bought a used Acom 1000 out of frustration since I have no other 6 M amplifier.

Thanks very much for taking the time to read all this. Suggestions are welcome, but I'm likely to say "tried that already" to any further suggestions. And, no, I do not have a dummy load that can handle 1 kW, and can't stomach spending several hundred dollars to buy one. Given that my other amps and radios see a 1:1 match on 6 M (the Acom reads "0 W reflected power", as does the Bird 43 and my other two new SWR meters), I don't think it will help.

73,

Eric NC6K

Poway, CA (DM13)

[Jeff DePolo]

Suggestions and comments:

1. Look at the bias supply output with a scope, make sure it isn't changing
as RF is applied and that it isn't being AM'med.

2. DON'T trust clamp-on ammeters in high-RF environments such as a Larcan
PA laying "on its back" for service with no shielding. Don't trust DMM's
either in such cases. An analog meter and external current shunt are your
friends here.

3. DO Measure the ripple/AC of the 50V supply with a scope while operating
the amplifier.

4. Operate into a dummy load instead of antenna to rule out possible RFI
issues.

5. Look at the output of the PA, confirm there are no parasitic
oscillations or other anomalies.

6. What are you using as an exciter, and have you tried an alternative?

7. Ferrites are going to do little to prevent low-frequency modulation of
the power supply rails, as might be the case if you have a parasitic
oscillation in the AF/ultrasonic range. Keep this in mind with regard to
#2, #3, and #5.

--- Jeff WN3A


---
This email has been checked for viruses by Avast antivirus software.
https://www.avast.com/antivirus

[eric gruff]

Thanks very much, Jeff! After posting my original message and before I saw your reply, I added a half-dozen Type 43 toroids and clamp-on ferrites to the DC line. I can now get 1100 W out of the Larcan before the B+ voltage rises enough to shut off the PSU (it is still exceeding 66 V, which means that the PSU is thinking the voltage is dropping and trying to compensate). I have never gotten more than 300 W out, so this is a huge step.

1. Bias is holding very steady at 38.9 to 39.1 V. B+ rises as output rises. It's not an artefact, and is causing two separate PSUs to shut off (blinking green LED that can only be reset by opening/closing the appropriate jumpers on the back of the PSU).

2. Agree. I fried my only shunt ammeter that can measure >10 A, so have to buy another one.

3. Dumb question, but how to interface my scope to 240 VAC without an explosion?

4. I don't have a dummy load that can take 1 kW, and really don't want to spend several hundred dollars on one. I may have to punt on this one.

5. I have Bird 40 dB sampling slug for 50 MHz. This will bring 1 kW to 100 mW, so I think I'll have to further attenuate the signal into the scope.What is the output of the sampling port on the Larcan? I see it's a stripline coupler, but am not sure of the attenuation.

6. I have multiple exciters (Yaesu FTdx5000, Flex 6300) that run through an external 6 dB attenuator before feeding the amp. I have a wattmeter in between to carefully measure what is going into the Larcan.

7. My issues have all been with only CW applied, so not sure if your comment applies to my issue.

QUESTION: Could I have fried a bypass capacitor that might result in more oscillation than usual? Seems like no one else is having nearly the RFI issues that I am with the same setup. My antenna is over 150' from the shack and at 60', but of course at 1 kW, that's still not very far. I do suspect the RFI is coming from the amp directly and not back from the antenna.

This is clearly an RFI problem, and I have carefully read your responses since I've identified, but not fully fixed the issue. I am going to bond the chassis of the Larcan and PSU with a braided strap (can't hurt), and will also add toroids to the AC input

I also suspect that my SO-239 output connector was poorly constructed and added to the amp (operator error by yours truly), so will replace it. 

Thanks again for taking the time to help me out. I feel better now that I have a good idea of the problem. Now to fix it.

73 - Eric NC6K

[Jeff DePolo]

> 3. Dumb question, but how to interface my scope to 240 VAC
> without an explosion?

No, I was speaking of the DC output from the supply. Look at it to make
sure there is no AC on it, which could occur if the PA were oscillating, or
if the power supply was losing regulation due to RFI, and that the DC level
is truly rising; it could be that the DC is staying constant, but there is
AC riding on it.

> 5. I have Bird 40 dB sampling slug for 50 MHz. This will
> bring 1 kW to 100 mW, so I think I'll have to further
> attenuate the signal into the scope.

I was suggesting that you look at the output with a spectrum analyzer, not a
scope. Guess I should have been more clear.

> What is the output of the
> sampling port on the Larcan? I see it's a stripline coupler,
> but am not sure of the attenuation.

I know I've measured it at some point, but I don't remember. It will vary
somewhat with frequency in any event.

> 7. My issues have all been with only CW applied, so not sure
> if your comment applies to my issue.

I was assuming you were using CW for testing. If there is a parasitic
oscillation, it typically results in AM components in the output signal.

> QUESTION: Could I have fried a bypass capacitor that might
> result in more oscillation than usual?

Possibly. Aside from frying the ammeter, what other problems occurred in
your early testing that lead you to believe something may have suffered
damage?

> This is clearly an RFI problem, and I have carefully read
> your responses since I've identified, but not fully fixed the
> issue. I am going to bond the chassis of the Larcan and PSU
> with a braided strap (can't hurt), and will also add toroids

> to the AC input.

I'd still want to take a look at the output from the amplifier before
agreeing that it's an RF ingress problem. If I were in your shoes, I'd beg,
borrow, or steal a good dummy load and a spectrum analyzer.

[Peter Sumner]

Hi Eric,

 how long are the control leads you have coming off of the PINS on the HP PSU?  are they SHORT jumpers or have you put long leads onto the control pins for a remote switch?   You made mention of using some sort of switch to reset the PSU, now these pins also have voltage sense and other functions and there is a possibility you have some RF getting back into this part of the PSU via any leads you may have soldered to the HP PSU's control pins.  I have just bent the pins on mine to enable the PSU and have no wires at all.  When I had wires the radiated noise from these short leads was quite loud, so in reverse any leads could pick up RF and be fooling the PSU into a fault condition.   Oh 1000w is prety good for a larcan on CW so be proud of that BUT be careful the FET;s are not very hardy when they get lightly offended :-)

maybe give this a go... just bridge the pins and see if that improves things before going back to any leads and switches.

Regards,

Peter, vk5pj, a proud LARCAN user in Australia

--
You received this message because you are subscribed to the Google Groups "Larcan 6m Amplifiers" group.
To unsubscribe from this group and stop receiving emails from it, send an email to larcan-6m-amplifiers+unsub...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[eric gruff]

Peter,

Many thanks! Yes, that was part of the problem. I pulled off the 3' long wire (which did have a small toroid that I wound it around twice - pretty lame, I know), and added snap-on ferrites to the AC input lines individually whilst I wait for some large Type 43 toroids to arrive. I also bonded the cases of the amp and PSU with grounding braid and I can now get full output with the PSU only rising about 10% (from 51.2 V to 56.9 V). This is a huge improvement over previously where I hit 70 V before the PSU faulted and shut off.

I can even bump the amp up to about 1200 W out, but haven't put the cooling fans back on and don't want to push my luck.

A big thanks to everyone who offered suggestions! You guys rock.

73 - Eric NC6K