Thursday, 6 June 2019

IC-7300, split IF, RF attenuator, transverter and ALC power spike

Now there is a complex title, designed to pick up every internet search ever invented.

Usual warning. This is a modification to a commercial radio, the Icom IC-7300. This may invalidate any warranty, so do not do it unless you appreciate the risks.

Usual assessment. The level of technical difficulty of this modification is, even by my standard, " very low".

The Problem
When I bought my ME2HT-Pro 2 metre band transverter I could see that I had found a superb performer. The thinking went that a device designed to work at maximum performance on a single band should be better than a multi-band rig. And so it proved. It really is superb. The only doubt I have surrounds its frequency stability while subjected to temperature changes generated on transmit. This issue needs to be resolved as I bought the high-stability version to avoid any trouble. It could be one of those snags which only exists in my head and not in reality.

I ran it initially with my Kenwood TS-590SG, but later I got an Icom IC-7300 which looked like a great combination. Gabi, who makes the transverter, suggests it as a option. Then a real-world problem arose because the IC-7300 puts out an RF power spike when going to transmit. This is common to many rigs as almost all use the ALC circuit to reduce the power. In the moment the radio goes to tx there is no modulation and the ALC is not active, leading to a momentary spike. Potentially this could be full power, which for the IC-7300 on 14MHz would be 100W, but in reality is lower than this. Another problem is that the spike is of very short duration, so it is very difficult to measure or even notice that it is there.

Anyway, undaunted, I connected the IC-7300 up to the transverter. I screwed the RF power output down to 1%. Every time I went to tx the output power LEDs on the transverter went full scale. Clearly these LEDs react to a short spike when my other meters don't. With the transverter producing 25W as shown on the LEDs (it is capable of 50W), the spike was lighting up the whole scale up to the 60W red one at the end of the scale. Scary.

With the transverter configured to 5 watt maximum input but adjusted to run to about 50% power on around 1 watt, this means that the spike must be more than 2 watts and quite possibly more than 5 watts. It proved too short to show up on conventional meters. I could only assume that it was in the 5 watt region or more. Recently I read a suggestion that it peaks at 25W.

This just compounded my often-stated reluctance to operate anything like a transverter on a single IF lead. All that has to happen is for the PTT line to fail, perhaps through tarnish on the plugs, and the TX transmits into the sensitive and expensive receive input section. Result: I would blow the whole thing up. Not that in 40 years of using transverters I have ever blown one up, but Mr Murphy would ensure that if I ever did it would be my best and most expensive one that went arrrggghhh.

The solution
Finding a solution to this problem involved a lot of debate with David, GM4JJJ, in the months before he sadly became silent key. The various ideas always involved splitting the rx line off from the tx line inside the IC-7300 (relays are too slow), ideally in a reversible operation. Then some way needed to be found to reduce the output power of the IC-7300. These two elements, tx and rx, would then be connected separately to the transverter in the same fashion as the TS590 used to be. But how to reduce the output power?

Well, those nice people at Kuhne Electronics make a board but I was disinclined to use it. There have been reported cases of PA instability and self-oscillation. I am sure that in most cases it works fine, but I also thought that the power handling of the resistor looked low and the soldering looked fiddly plus I don't like phono plugs for RF and then you have to take the board out to return the rig to normal use and ... I just don't like it. DB6NT's products from Kuhne are no doubt superb, but once again I am looking elsewhere.

INRAD make a simple kit for the IC-7300 to split the rx from the tx. It also uses phono sockets and costs £50 plus postage from major UK amateur stores. A bit more searching revealed a better one on eBay with SMA sockets and at a price which worked out at £22 with free postage. Not a difficult decision to make.
{EDIT} I hear in May 2020 that the eBay link above may no longer be available. Not sure if the product still is, but the INRAD one is currently still availble in the UK from Nevada and possibly other sources.
I thought about various ways to reduce the IC-7300s output power by isolating the PA. David doubted that this was practical, and eventually I realised he was right. He usually was.

I have finally decided that the simple thing to do is to reduce the output power with an attenuator. I did not go for that immediately as I associated RF attenuators with calibrated products which come at enormous prices. This is true - my first checks revealed that a 20dB 100W attenuator costs about £500, putting the whole project beyond economic viability.  Also, the range of power and attenuation is rather limited.

However, I then checked Amazon and found a shed-load of lower spec attenuators for £30 or less. There were cheap ones with few details, but also a mid-priced one direct from the manufacturer. Tracing back from the Amazon photo, I found the manufacturers website and even found something approximating to a data sheet. Suddenly, thanks to Amazon, there was lots of choice, no need to settle for a limited range of values available in the calibrated ones.

I did some calculations based on my assumptions about the power spike and the attenuator network in the transverter, and this came up with 10W for the spike which would need a 10dB attenuation to render it insignificant. I decided to go for a 25W 10dB attenuator. I doubt if the spike is 25W, but that seemed the best power/price option with a safety allowance for my guesses.

Installing the components
It wasn't difficult. The "RX Adapter" from eBay arrived from France quickly. I needed some support from the seller Daniel who could not have been nicer or more helpful.
IC-7300 RX Adapter from "hifi1200" eBay
There were no instructions but it is pretty easy to install. Just to be on the safe side I downloaded and checked the INRAD instructions - I'm not proud. All you do is to take the top covers off the rig (14 screws, including the ones round the speaker), pull the ATU socket out and turn it safely back inside the rig, replacing it with the adapter board. Then pull out the RX plug from the rig and connect it into the back of the adapter and then plug the adapter's plug into the empty socket. Clip on the filter. Then put the covers back on and test.
Adapter installed with the ATU socket twisted back out of harm's way.

The kit includes an SMA to SMA patch lead which returns the IC-7300 to normal operation. This makes it an easily reversible modification. After checking that this lead worked (it did) I took the patch lead off again and connected the rx feed from the transverter and the IC-7300 was working perfectly on 2m. The only slight snag is that the SMA sockets are a bit wider than the ATU socket was so the top cover of the IC-7300 shows a very slight bulge over the sockets. Not a problem for me as I don't look at the back of it much.
The neat board showing that the nuts round the SMA sockets are wide. They slightly bend the covers when fitted.

Moving on to the transmit side, the "RX Adapter" modification changes nothing. Thus the full power still comes out of the SO-239 socket on the back of the rig. The attenuator was being sent from China and took a while to arrive. Actually, it was just over a week! During that time the RSGB "Rad Com" arrived and in it was an article for a satellite transverter-type thing which also used a similar attenuator. In the article the author, Giles Read warned against connecting these devices up "back-to-front" as they have much lower power handling that way round (I wonder why). He said that they had an N-type plug which could be connected to the rig at the input end. I checked all this out with Giles who was very helpful. I think I have it the right way round.
RF attenuator from Amazon.
The device duly arrived with no instructions so the RadCom article had informed me as to which way round it should go and the danger of getting it wrong. I am never keen on screwing anything heavy like an attenuator directly into the SO-239 socket on the back of the radio. The weight will tend to distort the socket. Anyway, it would need an adapter or for me to shell out money to convert to an N-type socket on the IC-7300, only to hang a heavy weight on the end of it.

So guided by the Rad Com article I spent the time waiting for the packet to arrive from China by making up an unusual lead - PL259 plug at one end and an N-type in-line socket at the other. Surprisingly, I did not have one of those leads in my spares box. I hate plug sense RF adapters anyway.
PL-259 plug to N-type socket lead - you know you want one - and the attenuator.
By making the lead I did the only soldering required for this modification.

When eventually set-up with all the leads and the separate rx and tx, the transverter was still set to 5W from previous use with the IC-7300 in unmodified form. So I decided to try this and adjust the sensitivity using the two pots in the transverter.

The present order of things is that the IC-7300 RF output is set to 25%, notionally 25W. Very little output power is showing on the IC-7300 PO meter because output is limited by the mic gain for SSB or PC drive for data. This produces around 9W out of the IC-7300 on 14MHz.

The spike power is not adjusted by the power output control on the rig, so the object of the exercise is to reduce whatever it is by the same amount as the wanted signal which is influenced by the output control, and then re-balance them.

Trying to measure the attenuation accurately is beyond me. I do not have calibrated meters. The figures suggest that it is around 6dB but it could be more. Given the limitations of the meters I have then it could well be 10dB, but I wouldn't buy one from Amazon expecting to use it for measurement purposes. Spend £500 on a proper one if you want to do that.

The RX Adapter has split the rx and tx leads meaning the risk of blowing up the transverter receiver has gone. The attenuator has reduced the power into the transverter to around 2.5W. The spike is no longer visible in the transverter output LEDs.

The introduction of the attenuator is to reduce both the spike and the wanted signal by a fixed amount. So, assuming that before the mod the spike was 10W relative to an maximum input signal for the transverter of 5W. After the mod the spike should now be reduced by the attenuator to about 1 - 2W (depending on whether it is fully 10dB or the 6dB I think it might be), well below the maximum input power. I have adjusted the wanted output from the rig upwards to show 2.5W at the transverter. So the wanted signal should now be more than the spike, which the LEDs seem to show.

In reality I don't know what the power of the spike is, but I used to be able to see it higher than the wanted signal in the LED meter, and now I cannot see it at all.

There is no sign of the attenuator getting hot, but let us wait for an opening or a long Earth-Moon-Earth session to establish that for sure. I am pretty sure that the spike is so short that there won't be much heat in it, just enough voltage to do damage with the old set-up.

So far so good. Time will tell how it is judged in the long run.

I really like the IC-7300. It is not designed to work with a transverter, hence the modications. It does not feature to alter the readout for a transverter, which the TS-590 had. However, the TS-590 one didn't work very well anyway so I had it set to the IF frequency in any case. So either way it says 14.174 rather than 144.174. I can live with that.
IC-7300 on 2 metres, with 110hz correction checked against a GPS locked beacon. Not much happening.
Other thoughts.
This plan was hatched before the appearance of the IC-9700. I suspect it is a match for the transverter on 2m. It would make a great 2m radio for me. However, I have the IC-7300 and the transverter, so I should use them. Do not rule out the possibility that I might change to a VHF rig eventually, if only to get rid of the rats nest of wires.

It was also hatched before I suddenly became the owner of a second IC-7100, acquired in the SK sale. I didn't really need it. It works well enough on 2m, but not as well as the 7300+transverter combination. Still really good for 70cms where I have nothing to compare it with.

You should hear it. Relays in the rig (now not switching the RF), the transverter, and two in the linear, all sequenced and clicking in order one after the other, and then in reverse after tx, plus three fans independently doing their thing. Bang, crash, tinkle every 15 seconds. It sounds like it used to do in Knockmore Junction Signal Cabin when we signalled the Derry Fitted Goods off the Antrim Branch. That was what railways in 1967 were like, but not any more.

I hope that David, GM4JJJ, would have approved.





  1. Jim,
    I have successfully installed one of those RX adapters in my IC-703, and I plan do do the IC-7300 mod, too. Yes, I found the French one, too.
    Regarding the attenuator, I may make use of some 50W 50 ohms loads (adding a heat sink) I have, and add a relatively high value resistor for high attenuation to a transverter input.
    I see that we have similar ideas in that respect.

    73, Jan, OZ9QV

  2. Jan
    Yes, they say that great minds think alike. I think the heatsink is needed. However, the Kuhne board does not seem to have one, or perhaps I do not understand it.

    I had thought that it takes out the tx line before the PA. If the PA is still powered but not driven, I am not surprised that it can self-oscillate. So, I decided to go for the attenuator method.
    I remember in the 1970s seeing someone using a valve tx directly into a dummy load at 100W and taking a very low level feed from the dummy load. He did this by having the dummy load in a cage (KW Electronics dummy loads were like that) and then taking a very low level drive via induced RF from a wire he ran inside the cage. Others used a tap on the dummy load.
    Thanks for that and good luck. 73 Jim

  3. Jim
    When I built my first 2m transverter I was using a HW-100. There I removed the anode voltage from the PA and used the driver output. No problem with self oscillation, then ;)

    Now I do have an older FT-817 with a blown PA. That I might modify for use with a transverter, and then use a RTL-SDR as a spectrum scope.

    BTW on the IC-703 I had to file a bit off the adapter, then it all looked good enough for me to use. At the moment I am using the 703 as 10m TRX and having two other receivers (signal splitter) connected for monitoring, then I can still use the 703 for 10m (or transverter) and monitor the FT8 frequency and 28.200 for the international beacon net monitoring - at least when I do not transmit on 10m. If I need more power, I can add a 200-250W PA (that I have), or I can use my 100W HF transceiver.
    Good luck to you, too. 73, Jan

  4. Jan
    I remember the HW-100 but I never owned one. The guy with the dummy load was using a KW-2000. I never liked them much, very heavy too. Even when I was first into radio in the late 1960s KW gear seemed very old fashioned.
    Yes, I used an FT-101 with the valve-based plug removed from the back.
    Your idea for the FT0817 is good.
    The split rx/tx could be very useful in the future for several things. Although the 7300 has a waterfall, when I had the TS-590 I added a dedicated SDR to the transverter rx line and had a big screen display too. I used to use a Flex 1500 as transverter driver and it was very good too.
    If I could have disabled the power from the 7300 PA I might have gone that way, but it did not seem practical.
    I will be looking for you on the bands.
    73 Jim

  5. Jim,
    Do you know of an easy way to use the TX antenna on an IC7300 as a separate RX antenna feed when the 7300 was in the RX state? I would have thought it would have been part of this mod or the DB6NT mod, but they only appear to show the IC7300 RX disconnection and re-route of the RX input to a separate socket.
    Also the French Ebay link is now missing unfortunately, do you have any other contact details of the supplier?
    Thanks, 73
    John (G(3)PQA)

  6. John. As I understand it what you are doing with this device is putting a loop in the RX line between the tx/rx relay and the receiver. The loop comes out to the outside panel of the rig. On one side is a socket with the rx from the antenna, and on the other side is a socket which connects directly to the receiver. Thus you have the rx from the antenna as you describe on one socket. You could, for instance, put a splitter in the loop and feed two receivers. This would be 3dB down on each receiver, but you can get amplified and isolated splitters, or you could build one of those. Maybe 3dB doesn't matter anyway if you use lower bands.
    I am going to try to contact you direct about the product itself and if you do not hear form me please contact me QTHR (email address is on 73 Jim