Programming volatile memory boards in older Icom Rigs and radio upgrades

[AF6LJ]

For those who may want to play along at home...
Here is the service manual for the radio.
I would point out there are a number of undocumented items such as the first mixer diplexer that adjustment should not be attempted by those who are not sure what is going on. For whatever reason mention of this, and some tuning coils over on the PLL board were left out.
As I remember one of the undocumented adjustments involves setting the injection level of the PLL sub loop into the PLL mixer. Too much sub loop injection and you get spurs, too little and the loop becomes noisy.
When I bought my IC-745 I took it into work and ran a battery of tests on it.
which included spectral purity. After spending a little time with the PLL schematic I took the radio back in a month later and cleaned up the PLL a little.

These radios are getting long in the tooth, I bought my IC-745 as they were being discontinued in 1986.Thankfully I have a parts mule.

[AF6LJ]

I think, with a working RAM board and an empty one, it would be possible to make a piggyback adapter that allowed you to copy the contents of the good one to the bad one, no PC necessary.

Interested to know if there's a schematic of the board or at least a parts list (mainly just to know what the transistor is, I suspect a jellybean NPN type but...)

Would be a fun project to use something like a PIC or Arduino to replace the entire board, I think it would be easy enough to find a chip with enough NV storage that it could be a one chip solution.

The idea of replacing the logic board with a modern microcontroller has crossed my mind
You could also replace the matrix board, which not only is in a bad place on the radio but overly complicates the process of reading switches off the front panel.
Remote control could be added.

I do like the idea of cloning a known good Ram board...

Cloning looks like it might be very simple, just jumper the data and address lines between boards, assert WR on the board to be written to and count through the address space with a simple binary counter

Indeed, looks to be that simple.

[Ian.M]

The problem is 18 year old software with no sourcecode that bit-bangs the printer port to drive a 26 year old hardware interface design.

However, if you don't need to access the hardware interface it works OK under Windows NTVDM, and can be used to load, save and edit its .BIN files, which is sufficient to figure out the file format.   A very quick look indicates the file may be a simple sequence of 16 bit values, stored in low byte first order, of which the upper three nibbles of each are zero and the low nibble is the contents of that memory module location.

The easy way forward would be to use an Arduino for the memory module interface, and create and publish a sketch that can read and write ICOM memory modules, and transfer data to/from a PC, preferably in a human readable format e.g. ASCII hex dump or IntelHEX, so it can be driven from any terminal program and isn't tied to an OS or even to a particular interface between the PC and the Arduino or similar board.  An Arduino UNO should easily be capable of powering the memory module and controlling all its pins with no added hardware other than jumper wires.   The sketch should document the interconnections, and also use pin names rather than absolute numbers so it can easily be ported to future generations of the Arduino platform.   I believe an Arduino-based interface, with open source firmware will be highly future-proof, as Arduino-like boards are likely to be the de-facto standard for many years to come.

Ideally someone with a fully working rig, + a spare memory module that has already lost its contents would volunteer to do the development, as that would be minimum risk as they could test with the blank module and the BIN images to be found on the web, and confirm normal operation before risking their good module.   Otherwise, with a good reverse engineered schematic, + a N.O.S. RAM chip, it should be possible to build a functional clone of the memory board with a little effort.

You'll also need to write a utility to convert the existing .BIN module dump files to your chosen human readable file format.

The uPD444C 1024 x4 bit SRAM datasheet can be found at http://www.pisi.com.pl/piotr433/manuals/upd444.pdf

The rest of the info needed can be found in this thread, though it would be nice to actually ID that transistor, and get the value of the vertical resistor next to it, or to have a schematic of the RAM board to work from.   

[Radio Tech]

All, great information 

Ian.M
I am not the sharpest tool in the shed when it comes to this type of stuff. But I do have working radios and spare EX-314 boards. All are the 2 chip solution.  I will gladly sacrifice one. Will just need help on getting the correct data ect.

I will get some decent pics of the ram board and also try and provide a schematic of it.

[AF6LJ]

The problem is 18 year old software with no sourcecode that bit-bangs the printer port to drive a 26 year old hardware interface design.

However, if you don't need to access the hardware interface it works OK under Windows NTVDM, and can be used to load, save and edit its .BIN files, which is sufficient to figure out the file format.   A very quick look indicates the file may be a simple sequence of 16 bit values, stored in low byte first order, of which the upper three nibbles of each are zero and the low nibble is the contents of that memory module location.

The easy way forward would be to use an Arduino for the memory module interface, and create and publish a sketch that can read and write ICOM memory modules, and transfer data to/from a PC, preferably in a human readable format e.g. ASCII hex dump or IntelHEX, so it can be driven from any terminal program and isn't tied to an OS or even to a particular interface between the PC and the Arduino or similar board.  An Arduino UNO should easily be capable of powering the memory module and controlling all its pins with no added hardware other than jumper wires.   The sketch should document the interconnections, and also use pin names rather than absolute numbers so it can easily be ported to future generations of the Arduino platform.   I believe an Arduino-based interface, with open source firmware will be highly future-proof, as Arduino-like boards are likely to be the de-facto standard for many years to come.

Ideally someone with a fully working rig, + a spare memory module that has already lost its contents would volunteer to do the development, as that would be minimum risk as they could test with the blank module and the BIN images to be found on the web, and confirm normal operation before risking their good module.   Otherwise, with a good reverse engineered schematic, + a N.O.S. RAM chip, it should be possible to build a functional clone of the memory board with a little effort.

You'll also need to write a utility to convert the existing .BIN module dump files to your chosen human readable file format.

The uPD444C 1024 x4 bit SRAM datasheet can be found at http://www.pisi.com.pl/piotr433/manuals/upd444.pdf

The rest of the info needed can be found in this thread, though it would be nice to actually ID that transistor, and get the value of the vertical resistor next to it, or to have a schematic of the RAM board to work from.   

Good call...
A stand alone programmer could be cobbled together with a minimum of effort.

[CJay]

The vertical 'resistor' is a diode, it's probably something like a 1N4148 or equivalent, the transistor, as I mentioned earlier, I think it'll just be a jellybean NPN part.

The PCB diagram in the archive that has the 2Kx8 D446 SRAM in will be pretty much the same circuit, they're ignoring the upper 1024 byte chunk of the larger SRAM and they've also ignored the upper four bits of the data bus.

Probably worked out cheaper to fit the larger chip than source more of the older 4 bit devices.

If nobody beats me to it I'll draw up the schematic when I get home.

It will be pretty trivial to build up an adapter to plug one of those boards into an EPROM programmer, an adapter to write one in the same EPROM programmer would probably also be pretty simple.

Definitely wouldn't bother making a clone of the board as it stands, some form of NVRAM which retains contents even without power would be a far better idea, a 24C01 and a CPU would be a far better solution but I'm not sure about Arduino, can you manipulate the clock speed on them and disable the on board oscillator, I ask because the clock is smack bang in the middle of the HF spectrum and could cause all sorts of problems...

[Ian.M]

No, an Arduino or other MCU would *NOT* be a good choice for a replacement for the memory board.  Responding to an address pin change fast enough is likely to be difficult and adding an extra source of digital noise to a sensitive radio receiver is seldom a good idea.
   
While it may be worth considering a FRAM based replacement board, unless you need the extra memories that are provided by the commercial board that has been discussed earlier,  surely the easy option is simply to fit a battery holder (which will be a bit of a bodge job - Epoxy putty or 3M VHB tape + patch wire), which will get another 10 years out of these rigs before you have to revisit the problem.    Either way, you'll need some sort of programmer

[CJay]

No, an Arduino or other MCU would *NOT* be a good choice for a replacement for the memory board.  Responding to an address pin change fast enough is likely to be difficult and adding an extra source of digital noise to a sensitive radio receiver is seldom a good idea.
   
While it may be worth considering a FRAM based replacement board, unless you need the extra memories that are provided by the commercial board that has been discussed earlier,  surely the easy option is simply to fit a battery holder (which will be a bit of a bodge job - Epoxy putty or 3M VHB tape + patch wire), which will get another 10 years out of these rigs before you have to revisit the problem.    Either way, you'll need some sort of programmer

I don;t think the access speed is likely to be a massive problem, I am concerned that adding an extra clock signal into the mix, literally perhaps, might cause problems with the receiver but I think it's highly possible that an MCU could be run at the same clock speed as the main CPU, working on the assumption that Icom don't have issues with that 4MHz leaking into the receiver.

I don't have a radio to check out for myself but I can say that I've built PIC circuits before now for use in radios and experienced no problems with the clock affecting the receivers, YMMV of course and the IC-745 is a completely different beast to the ones I've played with, I think it's worth the experiment.

Adding a battery holder would still leave you open to losing the SRAM contents if you develop a bad contact or the battery goes flat so some sort of flash or non powered NVRAM is preferable.

[AF6LJ]

Replacing the Ram board with any kind of microcontroller is a bad idea.
If you are going to do that, replace the whole logic board, the Ram board and the radio's Matrix board. The signals on those boards are relatively low speed, in order to keep down the noise generated by all that digital logic.
Anything that runs too fast will be a PITA to suppress all the EMI...

[Radio Tech]

This is starting to get exciting.
Lets remember a few things as we move along.

1-At this present time we can order replacement boards with non volatile memory.  In fact I am going to order the IK2RND board for my own use.  I will not reverse engineer it as I do not want to take someone else work. Plus this is an option for those that do not want to experiment in the name of learning.  At this time however  how long will they continue to produce these boards?

2-Building a replacement board ourselves would be fun and exciting.  Should be a great learning experience for folks.  Will help us understand what was going through the minds of the designers during this time frame. My biggest thought was bean counters.  As Eeprom design cost way more back in that era.

3-For nostalgia,  learn how to use modern equipment to read and write to the existing EX-314 ram board.  For those like myself that would like to keep all the original parts intact and working. We already have the programmer and software designed to do this, what we do not have is the current delivery system.

I did dig out a win 3.1 and a win 98 pc from the back of the shop yesterday.  And they work I could simply find my old programmer or build a new one.  But how long these machines will run is beyond me.

Keep the thoughts and ideas flowing guys.

[CJay]

So could you make some measurements, if you could capture a read and a write sequence so we can get some kind of idea how fast they are?

Also would be interesting to see a capture of the control signals in relation to the data and address lines.

Plus, the number off that transistor?

It does occur to me that it may be possible to just fit or wire something like an Atmel AT28C16E in place of the SRAM chip but they're long imn the tooth and I don't know if they're still available...

[Radio Tech]

So could you make some measurements, if you could capture a read and a write sequence so we can get some kind of idea how fast they are?

Also would be interesting to see a capture of the control signals in relation to the data and address lines.

Plus, the number off that transistor?

This will be my priority for the next few days.  I been wanting a reason to learn how to use this logic analyzer. I will get some high res pictures and part numbers from the board this afternoon.  Will document and most likely record this journey.  Will post everything I find here.

[Silveruser]

I've been following along, Just replaced the battery in my R71A after all these years - I've no idea how many, maybe ten , maybe 20. I finally found the service manual after months of looking and found photo copies of these documents tucked inside. The first (attached) is a blurb from ICOM about how great this idea is. The second is a bunch of RAM dump printouts. I've no idea where they came from or if they are of any use but here they are from the days of dot matrix. If you think they would be of any value let me know - I'll see if I can find a way to share them.

[CJay]

The interesting thing in there is the bit that says the programming information is available from Icom and that users with a PC can experiment.

Which to me would suggest there's a way to load these in the radio...

Would be very interested to see more documentation.

[TheSteve]

The interesting thing in there is the bit that says the programming information is available from Icom and that users with a PC can experiment.

Which to me would suggest there's a way to load these in the radio...

Would be very interested to see more documentation.

They mention needing a home computer so I would think at some point Icom considered selling a programming interface.

[Silveruser]

They did make an Interface Unit EX309 for the radios. i have one installed in mine. Never used it. Don't recall ever seeing any instructions on commands/control possible. It just came with a single page with some pinout info and installation instructions.
It might be interesting if someone could find more of those Tech Talk notes. So far nothing much on ICOM sites, some more recent ones but nothing old. Maybe time to check the WAYBACK Machine.

[Radio Tech]

Silveruser,
that is great information. at this point we can use anything that is available for further testing. Thanks for sharing that.


I also agree that Icom may had that in mind after reading that article.  If they ever  let this information out is beyond me. I have contacted Icom in the past for information on old radios and all I got then was "Sorry, we no longer support that".  So unless we find a tech from the 80's that used to work for them I am pretty sure this information is mute. Hence why I started this thread here and started some videos on the 745.  I think I will also start hunting for "parts and donor" rigs to help with some of the experiments.  I do own a rig in good working order that appears to have the original battery in it. From that article the battery is only needed when the unit is not being used.

I just got home from work, went out to the shop and it is a chilling 38 degrees in there.  Will take an hour or so the heat up then I will spend a bit of time tonight looking things over. Really need to get the house lab back together now.

I do have  a EX-314 board in hand. That transistor (Q1) is a C945. Like you said cJay just a jellybean part.

D1,D2,and D3 looks to be 1N914.
C1 the big green ceramic is a 104M (0.1)
C2 is a 47 (47pf)
R1 47k
R2 390
R3 10k
R4 4.7k
R5 1.0k
R6 47k
W1 - W8 1 ohm
W9 jumper
IC1 HD7432P
IC2 D444C
TR1 C945

[Radio Tech]

They did make an Interface Unit EX309 for the radios. i have one installed in mine. Never used it. Don't recall ever seeing any instructions on commands/control possible. It just came with a single page with some pinout info and installation instructions.
It might be interesting if someone could find more of those Tech Talk notes. So far nothing much on ICOM sites, some more recent ones but nothing old. Maybe time to check the WAYBACK Machine.

Very interesting and thanks again. I did a quick search and found the below. I did not find it on ebay however, all have been sold.

[CJay]

Well that EX-309 seems to be an interface to the CT-10 which is capable of being interfaced to an RS-232 capable computer or terminal.

It's a bit of a mire of equipment.

I think tracing the EX-309 connections in the IC-751 logic board schematic, J10 and J15, might be interesting but we're diverging from the original subject now, I don't think it's of much use unless we get the CT-10 and whatever software Icom made available for that combo.

It's late here, I'm up again at 05:00 and it's almost 22:30 now so more digging tomorrow.

I reckon the parts list for the larger chip board will be exactly the same except for the D444 being swapped for a D446?

Silveruser,
that is great information. at this point we can use anything that is available for further testing. Thanks for sharing that.


I also agree that Icom may had that in mind after reading that article.  If they ever  let this information out is beyond me. I have contacted Icom in the past for information on old radios and all I got then was "Sorry, we no longer support that".  So unless we find a tech from the 80's that used to work for them I am pretty sure this information is mute. Hence why I started this thread here and started some videos on the 745.  I think I will also start hunting for "parts and donor" rigs to help with some of the experiments.  I do own a rig in good working order that appears to have the original battery in it. From that article the battery is only needed when the unit is not being used.

I just got home from work, went out to the shop and it is a chilling 38 degrees in there.  Will take an hour or so the heat up then I will spend a bit of time tonight looking things over. Really need to get the house lab back together now.

I do have  a EX-314 board in hand. That transistor (Q1) is a C945. Like you said cJay just a jellybean part.

D1,D2,and D3 looks to be 1N914.
C1 the big green ceramic is a 104M (0.1)
C2 is a 47 (47pf)
R1 47k
R2 390
R3 10k
R4 4.7k
R5 1.0k
R6 47k
W1 - W8 1 ohm
W9 jumper
IC1 HD7432P
IC2 D444C
TR1 C945

[Radio Tech]

Well that EX-309 seems to be an interface to the CT-10 which is capable of being interfaced to an RS-232 capable computer or terminal.

It's a bit of a mire of equipment.

I think tracing the EX-309 connections in the IC-751 logic board schematic, J10 and J15, might be interesting but we're diverging from the original subject now, I don't think it's of much use unless we get the CT-10 and whatever software Icom made available for that combo.

It's late here, I'm up again at 05:00 and it's almost 22:30 now so more digging tomorrow.

I reckon the parts list for the larger chip board will be exactly the same except for the D444 being swapped for a D446?

You may be right, was just thinking there could be some tid bits in the docs that could be useful.

I have attached the version A and B so the difference can be shown. I have the A version here somewhere.
Get some rest....

[Co6aka]

In the ZIP file for the "RAM Module Programming Utility" on "www.n2cbu.net" there are a couple of RAM dumps in both BIN and HEX; compare the two and you'll have the BIN format, which is pretty obvious.

PCB layout for the RAM board is in the ZIP, and attached here for simplicity.

EX-309 is just a parallel remote control interface along the lines of the serial CI-V interface.

[Radio Tech]

Thanks for that information. I have looked at the dumps of those but I am not a coder.  Will take me a while to figure all that out.

Any way back to what I was talking about on the radio programming a dead board by swapping it out while the unit is powered on....  Big fail. Will not work. As soon as the ram board is removed the display goes all "0". So there is something being sent to the cpu when the ram is there. I was 99.9 percent this would be the cause. Even Sue had her thoughts on it.  I recorded the experiment and will upload it shortly for your entertainment.

[Radio Tech]

The Failure 

[Co6aka]

As soon as the ram board is removed the display goes all "0"

The RAM board also contains/holds the current frequency, along with "VFO-B" and memories. I wonder if a supercap across the RAM chip's supply lines would hold it long enough the change the battery.

Anyway...best bet is to set up a PC with a good-old-fashioned parallel port and give it a go with the interface. I haven't programmed one of these boards since at least ten years, but it had to have been while running a P4 CPU. Key is the parallel port, not the CPU, so it should work with a PCI or PCI-X parallel port card, and those are pretty cheap. First grab a FreeDOS image, burn it, and see if you can boot your PC with it.

[Co6aka]

I while back a couple of Russians had done some RAM board hacking... So I poked around on Yandex and found this:

http://yl2rmk.qrz.ru/icom751a.html

http://www.qrz.ru/schemes/contribute/technology/icom751/english.shtml