HF Amplifier MX-P50M issues

I have a MX-P50M that has been moded to use with ICOM-705 radio.
I connected a power meter and dummy load to this setup.

For testing purposes, i am using RTTY mode on ICOM705 and the power meter is in peak power mode. The dummy load is rated for 50Watts, which is within the amplifier specs of 45W.

When the amp is turned off, the ICOM 705 working at 5W at 7MHz, i read 4.7W
When the amp is turned on, the selector knob on 40m, i read 5.4W on the power meter.
Here are the measurements from other bands
7MHz…push 5W on radio…4.7W (no amp) vs 5.4W (with amp) on power meter
14MHz…push 5W on radio…4.7W (no amp) vs 8.3W (with amp) on power meter
18MHz…push 5W on radio…4.7W (no amp) vs 7.5W (with amp) on power meter
21MHz…push 5W on radio…4.7W (no amp) vs 5.1W (with amp) on power meter
24MHz…push 5W on radio…4.7W (no amp) vs 4.6W (with amp) on power meter
28MHz…push 5W on radio…4.7W (no amp) vs 2.4W (with amp) on power meter

The battery voltage is about 13.9V, so that is not the issue.
The amp shows blue light, when turned on
The amp also shows red light, when i key the radio in RTTY mode

What am i missing here?
Any thoughts is appreciated.

Thank you
jim

1 Like

The amp has no ATU and requires a 50 ohm load. If you know the dummy load is reliable, the lack of output power suggests that one of the two chips comprising the final amplifier device has failed. From memory it is a Motorola device that would be obtainable by a radio service shop.

But… what about any other modes that produce constant carrier, such as CW or AM (half power usually). Just in case rtty mode is doing something odd (unlikely).

An unrelated question - did you have the amplifier modified to suppress negative voltage spikes when the ptt is “off”?

Andrew VK1DA

1 Like

It does sound like the final is broken, as Andrew says. However, the other possibility is a problem in the LPF (low-pass filter) section, where you switch between different sets of coils/capacitors when you change the band on the amplifier.
As you say, the amplifier has been “modded” for use with the IC705. Is this mod to enable the amp to switch bands based on the band selected on the 705? If so, the control cable between the IC705 and the P50M could be switching the LPF to the wrong band, which would cause this kind of problem.
A broken track or wire into or out of the LPF section could also be to blame. Has the amplifier been dropped, where a crack in a PCB could have resulted?
Another factor to measure is what current the P50M is drawing on the 13.8v line—if that is very low, it’s probably the final transistor block. If it’s high, then look at the LPF board and connectivity in general.

73 Ed.

1 Like

Hi Jim,

What were the mods? My first thoughts were if it was OK before the mods and not after then the mods are likely to be the problem. Check carefully.

My next thought was if the mod is OK then maybe the LP filter switch was out of register. So try 20 m for 40 m tx.

If these two possibilities are cleared then a dead dual FET semiconductor is most likely. Was the previous owner able to demo proper operation?

A faulty c/o relay might be also suspected.

Good luck.

73
Ron
VK3AFW

1 Like

Thank you for all the feedback and ideas.
Here’s some more data…i tried CW and i tried switching amp bands

Radio on 20m and Amp Switch on 20m.
(I am using CW mode on radio with a key)(Power meter is on peak power mode)
14MHz…push 5W on radio…11.7W peak power (with amp) on power meter
14MHz…push 6W on radio…18.1W peak power (with amp) on power meter
14MHz…push 7W on radio…26.2W peak power (with amp) on power meter
14MHz…push 8W on radio…31.1W peak power (with amp) on power meter
14MHz…push 9W on radio…33.4W peak power (with amp) on power meter
14MHz…push 10W on radio…35.2W peak power (with amp) on power meter
Testing out if the filters are at play: Radio on 20m and Amp Switch on 40m …Power meter shows no reading

Now repeat the test on 40m
7MHz…push 5W on radio…7.4W peak power (with amp) on power meter
7MHz…push 6W on radio…12.2W peak power (with amp) on power meter
7MHz…push 7W on radio…17.3W peak power (with amp) on power meter
7MHz…push 8W on radio…21.2W peak power (with amp) on power meter
7MHz…push 9W on radio…24.2W peak power (with amp) on power meter
7MHz…push 10W on radio…27.1W peak power (with amp) on power meter
Testing out if the filters are at play: Radio on 40m and Amp Switch on 20m …Power meter shows no reading

1 Like

So, my confusion is why am i getting amplified output on CW mode, well into 35W (albeit input at 10W) if the MRF186 is fried.
Here are some photos of the amp, nothing is hot or seems fried.

1 Like

Setup IC-705->MX-P50M->ATU-100->AV600 power meter->dummy load (1.1:1 or better from 1.8MH-1GHz) The AV-600 reads a few percent low on 80/40 but gets more accurate as the freq increases. ATU-100 power display is close enough but maybe a little more accurate. Both are ham quality, not R & S quality! The standard keying transistor mod has been done to the amp.

The 705 is in FM mode, RF power set to 30%, external 4S LiFePo
MX-P50 switched to correct band. Separate 4S LiFePo
AT-100 sharing battery with 705

Band ATU Drive ATU PA AV600 Drive AV600 PA
3.5MHz 2.9W 25W 2W 22W
7MHz 2.9W 25W 2.3W 27W
10MHz 2.9W 29W 2.5W 31W
14MHz 2.9W 26W 2.5W 30W
17MHz 2.9W 28W 2.5W 31W
21MHz 2.9W 28W 2.6W 31W
24MHz 2.9W 29W 2.6W 30W
28MHz 2.7W 28W 2.5W 31W

This appears to show the PA has a gain of 10 which fits. NXP claim the MRF-186 device has a guaranteed power gain of 11dB @960MHz @ 28V where the max output is 120W.

There is something wrong with yours. Before condemning the FET, I would confirm the band switch is switching the correct relays for the filters and also do some continuity checks / visual inspection of the coils etc. Also measure the current drawn by the amp, make sure it’s not being starved of power. Other than that I think you MRF-186 is somewhat sick.

1 Like

Hi Jim,
You didn’t say what the modifications were.

Something seems to be amiss with the filters if the 20 m filter shows nothing for 40 m input. They are normally low pass filters so set for 14 MHz it should pass frequencies up to at least 15 MHz. Set for 80 m and driven on 40 m will result in very little output and a protesting PA.

My Ameritron amp shows this characteristic. Setting it for the correct band eliminates some harmonics that would otherwise escape.

Your amp is horribly non linear with the gain increasing with drive. That’s not normal.

Half your dual FET PA is likely fried.

Mine gave a gain of x9 at 5 W in and x10 at half that input.

73
Ron
VK3AFW

1 Like

AB1PJ - Hello Jim.

MX-P50M HF 50W Amp / MRF186 RF Power Output Device Replacement

My MX-P50M Amp failed at the end of a CT3 activation. I thought at the time there’d been an antenna mismatch – maybe the wrong links were open in the dipole? Can’t remember.

Anyway my problem was the MRF186 RF power output device. A new one didn’t cost very much – £11 at the time 2017 then I bought two spares at £9 each plus a second MX-P50M amplifier in case it happened again. All from ebay.

The device consists of a flat copper plate with slots to accommodate screws at each end and a ceramic body. Two copper terminal tags protrude from each side of the ceramic body (five terminal device). They are quite simple to replace. Left to right, with ‘MRF186’ labelling facing you (so you can read it), I’m naming the terminals A & B along the top and C & D along the bottom (closest to you).

In case you need to, this is how I replaced it:

(NB: Take note of which way around it goes before starting)!

Unsoldered the capacitor across terminal tags A and B and removed for now.

TIP: I found it best to cut through the four terminal tags right up close to the ceramic body of the device using a sharp Stanley knife (new blade – 4 cuts). This was after messing around for too long with a 50W soldering iron and a solder sucker.

Removed the two securing screws.

Gently levered up the blown device using a flat bladed screwdriver. To avoid damage to the PCB, if it doesn’t lift readily, find out why!

Removed what was left of the old terminal tags from the PCB pads using the soldering iron and a sucker.

……

FITTING the New Device:

Cleaned everything

Ensured ALL solder was removed from the PCB pads so that the terminal tags on the new device sat down in intimate contact with its heatsink.

Checked for cleanliness again. (You don’t want grit in there). Applied a thin film of heatshink compound to the heatsink and a thin film to the base of the new device.

Using tweezers, placed the new device onto heatsink. Tightened the screws evenly. Excess compound squeezes out. Cleaned off.

Soldered each of the four terminal tags to the PCB pads in turn allowing for cooling between stages

Soldered the capacitor (removed earlier) across terminal tags A and B

……

TESTS on new and old devices:

Measurements using a DMM set to ‘diode drop’.

(no other connection to the device, no power or other stimulus applied):

NEW Device:

NEW Device with labelling orientated so it can be read:

Black DMM probe on base plate.

Red DMM probe on each terminal tag in turn:

All (A,B,C,D) open circuit (no conduction to base plate)

Red DMM probe on base plate.

Black DMM probe on each terminal tag in turn:

Top two pads (A and B): 0.57 V forward drop

Lower two pads (C & D): Open circuit (no conduction)

……

BLOWN Device (guidance only):

Black DMM probe on base plate.

Red DMM probe on each terminal tag in turn:

A: Short Cct

B: Open Cct

C: 0.014

D: 0.063

Red DMM probe on base plate.

Black DMM probe on each terminal tag in turn:

A: Short Cct

B: 0.49

C: 0.014

D: 0.063

Guidance only - this was what I found on my particular blown device with my particular Beckman DMM. Obviously other devices could exhibit all kinds of different readings depending on the circumstances of failure etc. However it is worth knowing what new and blown devices check out like.

Note: I always test on FM with a power meter and dummy load connected and taking quick readings.

……

Hope this helps.

Good luck

John G4YSS

Photos: 8-10-11-14-17


Above: Removing the capacitor


Above: Cut through the terminals with a sharp knife and remove the screws.


Above: Lift out the blown item


Above: Clean off and apply heatsink compound


Above: New device screwed down and ready for soldering in.


Above: Replaced

11 Likes

Great pics and description, John.

I think there is a perception among some owners that the amplifier can be band-switched remotely, ie. using band info from the driving transmitter. Mine can’t, it requires the operator to know how to use the amplifier band switch appropriately. If there is a mod available to decode band info from the driving transmitter and bypass the manual band switch it has escaped me.

On the other hand, I have read several versions of the negative ptt suppression mods various owners have made. I haven’t seen any of them described as an agreed standard mod. Is there one?

73 and hope the injury recovery is going steadily.

Andrew VK1DA/VK2DA

2 Likes

Thank you so much for this. I will get some MRF186 and perform this surgery. This seems very doable instead of discarding the amplifier.

2 Likes

Brilliant and clear guide John!

1 Like

Hi Jim and all. Thank you,

Before the MX-P50M came out I was using an MX-P50A which had the name ‘Sainsonic.’ It was sold by Amazon in 2013 for £154.59 and marketed under the name MX-P817. Despite that, it bears the name MX-P50A on its front panel, is a bit similar to the 50M inside and uses the same MRF186 output. (Amazon confused me as I’ve seen pictures of them with 817 labelling and no band switch).

More relevant to your point about the manual band switching, the MX-P50A has an ACC socket with supplied lead and the option of manual or auto band switching when plugged into an FT817. (The manual band switching option allows it to be used with other transceivers eg the IC703 etc).

The MX-P50A was a bit bigger and a bit heavier than I wanted for SOTA so, just like I did with my IC706-2G, I slimmed it down by making a composite lower panel (with foil screening re-added back) and a lighter alloy rear panel. Swapping the SO239’s for flying leads with BNC’s helped too. Even though I added a circuit breaker and rev. pol diode, the weight came down to just over a kilo - acceptable. The heatsink is bigger than needed too. I can’t remember it getting very hot, but I didn’t get around to butchering or replacing that, though I intended to. Anyway I digress (as usual!)

Though my Sainsonic MX-P50A (MX-P817) had served me well and reliably for SOTA, I abandoned it in favour of the lighter MX-P50M, which came along a bit later. I didn’t feel the need to modify the MX-P50M either. I doubt if you’d find a MX-P50A now but since they were sold by Amazon, there must be a few abandoned ones lurking in shacks if you could just persuade one out.

If in fact you did, you are not alone in getting the MX-P50M LPF selection wrong. Being used to the MX-P50A doing it for me, I went through half the activation of GM/SS-006 on 40m-CW before realising that my selector was in the 80m position resulting in ‘mouse power’ coming out. Thinking it was rotten band conditions I got mostly 339’s. That little episode didn’t blow the PA and I’ve done it a time or two since.

Product Description - MX-P50A (MX-P817)
Working freq: HF band 80M - 10M
Mode: SSB AM FM CW
RF input: 5W 50ohm
RF output: 50W
POWER input: 13.8V 7A
RF plug: SL16
SIZE: 20cm X 13cm X 5cm
Weight: 1.2kg

Quick Operating Manual (the English is a bit suspect!):

  1. Connect power cord to DC power supplier, Red to “+”, Black to" -"
  2. Connect ACC cable to FT-817 ACC and amp.
  3. Open the power on both power supplier and AMP. The power indicates on.
  4. Power on FT-817, the band indicate on Amp should be on.
  5. Connect antenna to amp “out” and “in” from FT-817 antenna out.
    Notice: The antenna you used that should well tuned, the SWR should less than 1:2.
  6. Press PTT, the TX indicate on Amp should be on.
    Notice: If working on FM or SSTV for long operating time, should lower the output power on FT-817, say 2.5W or 1W.

Package list:
1x MX-P817
1x Power cord
1x ACC control cable

“I think there is a perception among some owners that the MX-P50M amplifier can be band-switched remotely, ie. using band info from the driving transmitter.” Maybe this myth derived from the earlier MX-P50A described here?
Just a thought,
John

MX-P50A photos: 34/ 3-6-14:


Above: The Sainsonic MX-P50A sold by Amazon for the FT817/ FT817ND in 2013. They advertised it as MX-P817 (confusing)


The MX-P50A internal view. Some similarities to the MX-P50M but major differences too.


Above: MX-P50A rear view showing ACC socket for the FT817ND


Above: MX-P50A internal rear view of ACC socket

4 Likes

Well I stand corrected. Did not know there was an earlier model with auto band switching. The change may have been to widen the usage of the amp to other radios than the 817 series. Possibly triggered by the variety of qrp radios coming onto the market.

Thanks John.

73 Andrew VK1DA/VK2DA

2 Likes

Hi Andrew,

I’ve been concerned Jim has had band filter selection problems. The mod to a 705 connection is a concern as they use a different switching scheme to the 817. It requires an interpretation matrix and maybe a driver. It could be done in a small chip.

I asked twice but got no answer to the question of what were the mods. Maybe Jim doesn’t know the details.

It’s unclear how he selected the bands.

I haven’t asked if the seller said that amp was 100% operational. There are some who sell you something in a sack which isn’t as claimed. Silence usually means a problem.

I wish Jim luck.

I hope he is able to fix the problem.

73
Ron
VK3AFW.

5 Likes

Hi Ron
I am not sure what is the best way to test band selection problems.

Here’s what i tried:-
I pulled out the rotatory band selector knob and noticed there is a screw that can be rotated 4 different positions. For kicks, i selected each position and transmitted with amplifier on.
As i shared, in the above post, when using CW mode (also tried RTTY and FM mode) … Radio on 20m and Amp Switch on 40m …Power meter shows no reading. Also tried other bands, and got no reading. However, band selection on 20m, i get readings but the numbers are all over. I see no linear amplification.
I repeated this for 40m test and saw similar results.
My conclusion was, the PA is not doing its job.

1 Like

Hi Jim,

Thanks for the clarification. I agree, it’s not looking good for the PA chip.

I often use RTTY when I want a stable carrier at full output. That doesn’t work for all rigs. FM can do the same but of course if the mic is connected you get some incidental modulation. CW is ideal but you lose one hand to holding the key down.

I normally set the AM carrier to 25% if the rig doesn’t do it automatically. That’s useful for longer testing such as tuning an antenna matcher.

Was the mod you mentioned to use the 705 intended to auto band switch the output filters?

73
Ron
VK3AFW.

2 Likes

Ok, i was able to do a butchered job of de-soldering the PA and take some measurements in diode mode (like described above). It took me an hour or so using my Hakko FX-888 (Finn-ish brand) soldering station. I set the temperature control for the soldering iron at 400C. Also used solder sucker from Precista T2C (Swed-ish brand). I will have to do more cleanup using the solder sucker.

Here are the measurements :-

DMM Black probe on base: Red probe on A,B,C,D >> A (open),B (open),C(open),D (1179)
DMM Red probe on base: Black probe on A,B,C,D >> A (609), B(608),C(open),D(1180)


2 Likes

Sorry Jim,

Can’t decode your email. This device is a dual FET. FETs don’t have a base. They only have 3 terminals. Base plus ABCD doesn’t compute.

Your MM should show open circuit (OC) from Gate to either Source or Drain. There should be an OC to the adjacent transistor.

The Sources will be common and connected to the gold bar. The Gate is the input and the Drain the output.

There should be a high value resistance between Source and Drain. If it’s infinity it may indicate a blown internal interconnection.

Your FET might be OK but I can’t say as I’m not sure what you have measured.

I have a smart device tester - that is the only way I can tell if something is fizzed, well except for bipolar transistors where a MM is useful.

73
Ron
VK3AFW

2 Likes

Read John’s description. There’s the mounting flange aka “base” and four terminals A,B,C,D which correspond to terminals 1,2,3,4 on Motorola/Freescale/NXP’s data.
flange = source for both FETS
1 = A = gate 1
2 = B = gate 2
3 = C = drain 1
4 = D = drain 2

Happy to be corrected.

2 Likes