KA7OEI - Substituting the SRF7043

Substituting the SRF7043 in the Yaesu FT-100
(plus some info about the MRF1570)
Before undertaking any replacements described, please be certain to read the warnings at the bottom of this page.

A "relatively" common problem with the Yaesu FT-100 is the failure of the VHF/UHF final transistor. Exactly why this is is open for debate - and the causes may be multiple. An incomplete list of possible causes may include:

Operation of the transmitter into a poor load. Even though the transmitter has circuitry to protect against high SWR, these circuits - no matter which radio they may be in - are not foolproof and do not react instantaneously and damage to the output devices may still result.
Poor grounding on the LPF section. One suggested cause has to do inadequate grounding of the lowpass filter/TR board that could result in some sort of feedback and/or poor-quality load.
Idle current runaway. It has been suggested that residual flux left on the leads at the factory may result in the development of a leakage path on the device, causing idle current to go up - even when no external bias is supplied - thus overheating and destroying the device.
Poor selection of the output device. Some have surmised that the SRF7043 was not the proper choice for the output device in this application.
Poor quality of the output device. Yet others have suggested that the quality/ratings of the SRF7043 were not as good as the published values.

On the above points, I make no judgments and consider them all equally to be potential problems. To be sure, several of these points seemed to have been addressed in later production runs of the FT-100/D - but whether or not these modifications reduced the likelihood of device failure is unknown - to me, at least.

When my FT-100 (non-"D") version failed, it did so somewhat insidiously. Transmitting occasionally on 2 meter SSB into a reasonably good antenna (but not the best - the measured VSWR on 2 meters was 2:1) everything seemed to be going well - until I smelled something "hot" - and it turned out to be the FT-100. Later checking showed that the idle current of the transceiver was on the order of 10-20 amps - and fluctuated wildly. Another quick check showed that there was no longer any 2 meter or 70 cm output power.

Popping the cover, I quickly found the source: The VHF/UHF final transistor - the SRF7043. Investigation using a jeweler's loupe showed a tiny hairline crack in the ceramic case of the transistor - a sure sign that something was very wrong. A quick check of availability of this part also showed that I was in for sticker shock: A new transistor - if available - was typically on the order of $200 - more or less.

Trying to dig up data on the SRF7043, I noticed several things:

It is a push-pull "dual" transistor in one package.
The SRF7043 wasn't a very common device - probably a Yaesu-specific house number.
Reports (as of 9/2006) seem to indicate that Yaesu service has not been routinely replacing the SRF7043 in radios sent in for repair, but rather simply returning them, un-repaired, with an explanatory note that states that the part is no longer available. Although Yaesu is aware of other devices that can be substituted to provide at least some RF output on VHF/UHF, they have been unwilling to use them - probably due to a lack of an engineering recommendation from the parent company.
More recent reports (5/2007) imply that Yaesu has been declining to perform any service at all on FT-100's: If you have personal experience with this, please let me know about it! There are other reports that some third-party companies are still performing repairs on the FT-100, but only where parts are available.

Exactly what other device would substitute for an SRF7043 wasn't, but rummaging around on the net I found that PolyFet makes a number of similar devices.

This shows the installation of the SK702 in the FT-100. Note the removal of the capacitors across the output terminals of the transistor and the **reconfiguration** of the "output" side of the balun - see text.
**Click on the image for a larger version.**

It has been suggested by others that a possible substitute for the SRF7043 was the MRF166W, which is equivalent to the PolyFET SK702, so I ordered two, as the minimum order was $200 with each device costing just over $100 each when shipping was included - still cheaper than just ONE SRF7043 - if you could get it! Note: Contrary to other information, the MRF166W is NOT equivalent to the SRF7043!

Results:

Simply dropping the SK702, a 28 volt device, in place of the SRF7043 resulted in rather poor amplifier efficiency: At saturation, there was about 10 watts out on 2 meters with about 6 amps of Drain current and 5 watts on 70cm with about 4 amps of Drain current.

Removing all of the capacitors on the output of the device (namely C3553, C3555, C3607, C3608, and C3609) improved things somewhat: At 13 volts of drain voltage, the unit saturated at just over 30 watts of RF output with 6.5 amps of Drain current on 2 meters and almost 20 watts of RF output with the same current on 70cm.

Another recommended modification is one made to the coaxial 1:4 output transformer using the yellow coax cables. These cables (which are appear to be pieces of 17 ohm coax - yes, 17 ohms) transform the fairly low output impedance of the transistors' drain circuits up closer to 50 ohms and do so in a balanced manner - hence the "push-pull" circuit. While testing with the SK702, I observed that these coaxial transformers seemed to slightly affected by their routing and proximity to nearby components. To eliminate this tendency (which may be related to circuit instability - but this is only a guess) I reconfigured the transformers by unsoldering the "non-transistor" end completely, connecting the center conductors of the coax to the outputs as shown in the picture, then connecting the shields together, leaving them floating. After this was done, the 1:4 transformer appeared to be completely immune to routing of its coax. It is worth noting that the "new" configuration is functionally identical to the "old" one in terms of impedance transformation.

For the time being - and to preserve linearity of the output (extremely important if one operates CW, SSB or AM) I set the "Maximum" output power on the adjustment menu to provide a maximum of 15 watts output on 2 meters and 10 watts output on 70 cm - this, taking into account the need to be able to operate linearly with a power supply voltage of just 10.5 volts as might be expected with wire losses and using a weakening battery.

Comments:

While this device will work, it falls somewhat short of the original device's output capability - about 6 dB or 1 S-unit.

Trying the LK722's - an expensive experiment:

After first trying the SK702, I ordered a pair of LK722's to try out. In speaking with an engineer at PolyFET, it was determined that, for maximum power in a broadband amplifier circuit, this would be a better choice than the LK721 as it was designed for 12.5 volt operation and appeared to be more rugged. They were $104 each and with a $200 minimum invoice, I obtained two of them.

During the initial testing, it seemed to be fine in that I could obtain about 45 watts of saturated output power on 2 meters and about 30 watts on 70cm. Like the SK702, it, too, seemed to operate with best efficiency and power output without the output loading capacitors (C3553 etc.)

Everything seemed to be going fine: There appeared to be no obvious sign of instability and the devices seemed to be operating with good efficiency and were not getting unduly warm - and suddenly, the output power dropped by about 6 dB.

Investigation first revealed a loss of the gate bias supply, dropping the devices into Class-C operation and lowering their gain. Further investigation revealed that they had developed a low-resistance Gate-Source resistance - about 15-20 ohms - which swamped out the gate bias. All this time, I had been observing the output of the amplifier on the analyzer and I saw nothing that had caused concern, and output power versus drive seemed to be stable - a reasonable indicator of device stability.

Carefully re-checking the output circuitry for any obvious failure (and finding none) I put the 2nd device in: It, too, operated flawlessly for a few minutes before the same thing happened.

Hmmm... (Actually, I said other words - none of which I will repeat here!)

So, I put the original SK702 back in and it worked exactly as before. Being somewhat worried about ruggedness, I did several things to try to kill it such as unterminated/short the output at full output - but nothing seemed to faze the device, so I put everything back together again.

So what happened with the LK722's? I don't know, exactly. The most likely cause was some sort of spurious oscillation, but I saw no evidence of such in all of the testing. I will leave it up to someone braver than I to conduct further experiments along this line...

Additional Comment on the use of the LK722 or similar devices:

I have seen claims that an LK722 (or other devices) will simply "drop in" as a replacement and give full power output, but these sources, when asked for any specific details about the replacement procedure and about the performance, have not responded.

If you have successfully used the LK722 (or any other device) as a replacement for the SRF7043 (even if it is at lower power) please let me know, along with some details and your experiences so that they can be shared here!

After device installation:

Once the device has been installed, some modifications are recommended to the "Alignment Menu" items. These modification are required as you have reduced the amount of output power that the radio is capable of on VHF/UHF. Failure to do the steps involving F-21 through F-23 will likely result in a badly distorted and "splattery" signal on 2 meter and 70cm sideband.

These steps are:

Set menu item 23 to "100" for maximum output
Go to 2 meters (145 MHz) and select CW. Make sure that break-in, etc. is set so that closing the KEY will actually cause the transmitter to key.
Set alignment menu item F-21 (144 PO 20 W) so that the output power is 6 watts when keyed.
Set alignment menu item F-22 (144 PO 50 W) so that the output power is 15 watts when keyed.
Using the "F" button, set the meter so that it displays ALC.
Set alignment menu item F12 (V/UHF IC ALC) and key the transmitter: Set for 1 "ALC" dot.
Go to 70 cm (440 MHz) and select CW.
Set alignment menu item F-23 (420 PW 20 W) so that the output power is 10 watts when keyed.
Make sure that you follow the steps in the manual under "REV ALC Adjustment." Only those steps pertaining to VHF/UHF need to be performed. NOTE: It may be possible that 10/15 watts of RF is insufficient output power to trigger the "SWR" detector when the 16.3 ohm load is used. In this case, you may have to settle for a setting of "5" or so in the appropriate menu (e.g. "F-45" for 2 meters, "F-46" for 70cm.)
It is also recommended that the "TX Gain Adjustment" steps be performed for 2 meters and 70cm, adjusting for and output power of 12 and 8 watts for 2 meters and 70 cm respectively.

Datasheets:

SRF7043 - The original transistor used in the FT-100. (This multiply-copied data sheet is only partly in english, but the important parameters are easily read - Thanks to Bob J. for passing this along.

PolyFET LK721 and PolyFET LK722 - These are 12.5 volt devices: See comments on the LK722 in the sidebar.

PolyFET SK702 - This is the device that I tried and describe above. It works fine and seems to be quite stable and rugged - if you willing to settle for lower output power than with the original device.

PolyFET SK701 - This device will also work - but it produces LOWER power than the SK702 - see comments below.

Comments about the SK701:

The SK701 is often found on Ebay for a reasonable price and is often shown as a "Substitute" for the SRF7043. It should be noted that the SK701 is a LOWER POWER device than the SK702 mentioned here - a fact usually NOT mentioned by the Ebay sellers!.

In correspondence with others who have used the SK701 instead of the SK702 as a replacement, here are a few things that you can expect:

Only about 10 watts of RF "clean" RF output on 2 meters. Being that the SK701 is also a 28 volt device and somewhat "wimpier" than the SK702 - not to mention higher output impedance - it cannot deliver anywhere near its 45 watt rating at just 12 volts.
It will be necessary to modify the output matching transformer to obtain maximum power from the device. Even doing this, however, will still yield only 10-12 watts or so on VHF and 5-8 watts on UHF if one wants to preserve linearity and have sufficient margin for operating the radio on somewhat discharged batteries. If you NEVER use any other mode other than FM on VHF and UHF, more power is available, but linearity on CW, SSB, or AM will be unacceptable!
As is the case with the SK702, you will need to go into the configuration menus and adjust the settings so that the SWR protection, power control and ALC function properly to assure safe, linear operation.

MRF1570 Data Sheet:

Here is an annotated copy of the MRF1570 Data Sheet for your reference, with thanks to the person who sent it to me.

Page 1
Page 2
Page 3

Other possible replacement devices:

According to Bob, KC3VO, it is possible to use an MRF1570 device (after minor mechanical modifications, but little or no electrical modification) to replace the SRF7043 and achieve excellent power output and stability: For more information, contact him directly.

A note from Duje, 9A3DDG relates: " I just want to say that ... [the] (MRF1570) better works without output capacitors, 12-22-39-56-68pF. Also, i put it on spectral [analyzer] and on cw mod full power 433MHz adjust output bias for minimum spurious. It is a must. "

Thanks!

Warnings
Please read the following:

The author of this page can, in no way, guarantee the accuracy or completeness of this document. It is solely up to the reader to determine the suitability of the information presented here and whether or not it pertains to his/her application.
There is no guarantee that the substitution of/by the device(s) mentioned in this article will work in all cases - or in your case.
It is not recommended that the substitution in this document be performed unless the person doing so has appropriate technical experience, skill, and access to test equipment: If this procedure is done improperly, it is possible that damage to property or, possibly, personal injury may result.
Many RF power devices contain the highly toxic substance, Beryllium Oxide. While it is not known for certain whether any of the devices used in the FT-100 - or those suggested as replacements - contain this potentially dangerous substance, in the absence of certain knowledge, it is highly advisable to make the presumption that it is present. It is up to you to learn about and take appropriate precautions when handling RF devices, including - but not limited to - wearing face masks and/or respirators, gloves, proper disposal of defective devices, and proper cleanup and decontamination of the work area.

Go to the KA7OEI main page.

This page last updated 20181119

Since 12/2010: