EIP 545B Frenquency counter

[Dave Cawley]

Thanks HP, I'll try that later but we are entertaining Jill's Mum and a Church lady today !

BTW I missed the keyboard test page.

More later

Thanks

Dave

[Dave Cawley]

.
Well, I checked all the Tantalums and they were all just fine, apart from the one I already replaced that had a dead short.

I have cleaned all the interconnect pins between the two boards and with some new pins exercised the sockets with switch cleaner.

So far it is the same as before,  displays the correct frequency on Band 3 but unable to switch to any other band or to switch the power meter on, only the power meter offset and a frequency limit switch seems to do anything.

U12 pin 2 the keyboard interrupt stays low until any key is pressed.

U11 the MC68B21P, Figure 111a shows the clock on PA1 pin 3 to be continuous, I cant's see anything there.

Any help greatly appreciated.

Thanks  Dave

[Dave Cawley]

The clock waveform/drawing isn't scaled or accurate in the service manual.

It turned out to be :

I had another look, this time with my elderly but lovely HP Logic Probe.  When I probed the 74LS195 pin 10 the clock input, and this chip is the keyboard scanner, it started working and kept working.  I had a look and only on that chip nearest the top cover edge,  had corroded pins, so I put it in an IC runner/straighter and cleaned one side of all the pins with very fine emery as well as switch cleaning the IC socket that looked OK anyway.  It's now stable as a rock, even the limited power meter function works !

So just one tantalum capacitor and a clean of one IC's pins.

Thanks for the help and inspiration here.

Dave

[HighPrecision]

Sorry for late response, I'm living very sad days in this period, just lost my beloved father   

Good, I've suspected the KB row counter U3, well done.

From your first post I expect a 548A model, the PIA 68B21 is present in newer units (548B), based on the new A105 microprocessor board clocked @2MHz, check the A105 if it's a newer board, also installed in 548A CCN# 2309

[Dave Cawley]

.
.
Very sorry to hear about your father.

Yes, I said it was an "A" when in fact it is a "B", not sure why I said that.

The good news is that now it is perfectly responsive whereas years ago,  when it last worked,  it was often difficult to drive, maybe the corrosion had already started to set in.

Attached is a photo of the Micro board with 3 chips missing !  But it works and I believe it has never been touched as I know the original purchaser.  BTW if you right click and copy the image it is 3K wide.

Regards

Dave

[HighPrecision]

This board is the old type @1MHz clock and ancient EPROMs TMS2516/2532, empty sockets are for option firmware extensions, your Opt.02 Power-meter EPROM is located on A107 board.

New board is much different, see this link:
https://911components.myshopify.com/products/eip-micorwave-2020215-02-a-circuit-card-assembly

See my previous posts for more detailed explanations.

[Dave Cawley]

That's odd ?  I also have the band 4 option but not the external sensor, but I don't need it.

Do you think my counter may not be as original as I thought it was ?

Whatever,  it is doing what I want so I'm a happy bunny !

Regards

Dave

[HighPrecision]

That's odd ?

Not necessarily 

Your hardware configuration depend by CCN number on a label at rear panel, unless this unit was modified previously by others (board exchange), the presence of MC68(B)21P probably indicate a possible previous attempt to resolve the keyboard problem.

I have a 578B with some options (01,02,05,06) and this unit have some old boards, clearly exchanged with other units, but after repair it's fully operative with very good sensitivity on band 3.

[Dave Cawley]

  Just went into the workshop it is a 548A, options 02, 06 serial number 02063  ccn 2309 or maybe 2307 ?

Sensitivity on 2.7Ghz -32dBm

[HighPrecision]

Just went into the workshop it is a 548A, options 02, 06 serial number 02063  ccn 2309 or maybe 2307 ?

Sensitivity on 2.7Ghz -32dBm

A105 board in your photo is consistent with the CCN 2307

band 3 sensitivity at -32dBm is good as the majority of these units, some A203 MW converters exhibit more sens as mine at -37dBm.

[Dave Cawley]

That's good then.  The final thing is that the power meter will only function on Band 3.  Dosn't bother me though.

Thanks

Dave

[robots]

I am very sorry to open VERY old thread, but....

I have also bought EIP 545B from israel. 2 units. One was not working, the other blew fuse. First unit repaired itself on way to me - probably just bad connection on the interconnect. Second had transformer shortage.

But back on topic. Power meter is not enabled on any of these units and both have that special WB firmware. I put the firmware into Ghidra (there is support for 6805 cpu in git).

After day of digging:

eeprom offsets:
0 - set to 0x00 to enable power meter option
2 - set to 0x5a to enable DAC option
3 and 4 set to 0x35 and 0x59 for some magic (probably band 4 operation, I have not traced this, not sure what it does)

Frequency offsets:
As "douardda" found out, the frequency offset is set in firmware right at the beginning, together with resolution.
Frequency offset can be disabled by removing the code, led will stay light up.

for resolution:
This is the code from douardda's blog

Code: [Select]

6125             LDX     #M005D                   ;8E 00 5D
6128             LDA     #$0D                     ;86 0D
612A             STA     $01,X                    ;A7 01
612C             LDA     #$05                     ;86 05
612E             STA     >M0045                   ;B7 00 45
There is however very important function called right after.

Code: [Select]

6131             JSR     Z66FB                    ;BD 66 FB
This function will commit the resolution value to various places, probably for counter chain. I have traced the gpib commands and from there i have found how resolution is set.

So to correctly set resolution to full scale, write 0 to it:

Code: [Select]

612C             LDA     #$00                     ;86 00
612E             STA     >M0045                   ;B7 00 45
6131             JSR     Z66FB                    ;BD 66 FB

I still dont have idea what that 0x5d is. If someone wants i can share my ghidra project files. My firmware is newer than douardda's. I have no idea what is newer.

I have dumped firmware from both units. Both firmwares are identical, except for EEPROM. Get them here if you want: http://robot.mysteria.cz/eip.zip
None of these findings have been verified on real HW yet.

[robots]

To enable power meter press:
TEST 1 0 - to enter test
0 8 0 0 - to read eeprom address 0, on display you should see content of the memory as last 2 digits
0 0 - to write eeprom address 0 to 0

To enter test 10, you need to flip the switch on band 3 converter board. If not you will get ERROR 20.

I had to repair my VCO board, which is quite different from the one posted on page 2. My Microwave assembly is -9 REV C, counter build year is 1995. This seems to be very new unit.

http://robot.mysteria.cz/IMG_20200210_103629.jpg

Problem was burnt resistor R25 - 51.1 ohm - almost in the center of the board. This resistor feeds power and limits current to the VCO. VCO is powered from 12V line. Resistor burned because one ceramic capacitor was shorted to ground (0.8Ohm measured). These ceramic capacitors are marked(!) AA5 which translates to 0.1uF.

Quick reverse of this board:
bottom left to right: tuning of the VCO, R6+R9 limit current, Diode D2 (marked A0) - reverse protection shotky, capacitor and inductor to filter the signal.
D3+D4 are M4A varicaps.
Q2 is probably the vco transistor, and Q4 is buffer transistor/amplifier, signal is divided into 3 parts on the "red" resistors.
U3 and U6 are signal amplifiers for the big-ass transistor BLU11/SL (2.5W).

top left: 5pole low pass filter and U2 (same as U3) MAR-8 MMIC amplifier.
U1 is not populated, probably the same transistor, and its the third output from the VCO. Probably for band 4 operation.

Replacing the capacitor and resistor with something i had (0805 100nF and 2x100ohm 1206 in parallel) and the vco woked! This brought band 1 and band 2 to life. Problem now is band 3 not being able to lock on frequency. Replacing Gate board from second unit did not help. I will need to dig into the IF part of the microwave unit.

[robots]

VCO part repair:
VCO output (one that goes to band 2 converted) should be about 0-1dBm. If not the U2 is probably faulty.

VCO amplifier is 3 stage amplifier:
u3 - MAR-8, Q6 - rf bjt, and the big transistor mounted on the body.

Enable VCO amp (test 1 0, 1 8 4 3, 3 f(reset)).

Output after u3 should be about +7dBm. Output of the BJT transistor should be about +17dBm.

If VCO amp is disabled (test 10, 1843, 37) last 2 stages are turned off, and but U3 stays enabled.

In my case U3 was faulty. As MAR-8 is very hard to obtain I went for easier to obtain RF amplifier - SPF5043Z.  I replaced U3 with U2. This way the VCO amplifier is in the same specs(or similar) as before. And i placed SPF5043z in U2 position.

SPF5043Z is bit different: Input voltage is max 5Volt. I placed little LDO in SOT23-5 package on free ground plane. Wired 12volts to input, and output directly into the L7 inductor through 0ohm resistor. R14 (100ohm resistor) was removed.

I also had to replace R32 (200ohm resistor) with 50ohm, to get the signal strength a bit higher. At -6dBm output the counter selftest (test 01) would not count 200MHz.

Here is picture of repaired VCO output:
http://robot.mysteria.cz/IMG-20200220-WA0002.jpeg
Mess is just flux residue, ground planes are hard to solder, as they are placed on the aluminum case.

I tested the counter to be mostly in specs. -30dBm signal at 6GHz through 1 meter of RG316(~ -2dB) + 2x N to sma (~ -2dB?)  ~36dBm sensitivity.


Second counter would not lock VCO to correct frequency. It turned out to be MC12013 on the converted control board. These ECL logic circuits are almost unobtainable (china virus stopped postal service) or for very high price (>20eur + s/h). I suggest scrapping from old equipment. I scrapped mine from 14GHz SSPA from the same era. MC13013 with MC10131 form 40/41 dual modulus prescaler.   If someone knows some 40/41 dual modulus prescaler which can go up to 550 MHz please let me know. (SP8793 or MC12016 are not good alternative)

I hope this will help someone in future :-).

[robots]

eeprom offsets:
0 - set to 0x00 to enable power meter option
2 - set to 0x5a to enable DAC option
3 and 4 set to 0x35 and 0x59 for some magic (probably band 4 operation, I have not traced this, not sure what it does)

eeprom offset 3 (address 0x803) - if value here is 0x35, counter counts to 26.5GHz , if other, 20.5GHz. Could someone confirm this ? i have no >6ghz source.
eeprom offset 4 (address 0x804) - if value is 0x59, band 4 is present

Main difference between firmware 2060048 and 2060052 i have noticed is the GPIB board detection and gpib_present flag.

[douardda]

@robots I've been away from the forum for quiite a while so I missed your execellent posts, very nice.

I think I'll get back on the last bits of tuning my EIP firmware when I'm finished with my current replacement HP34970A front panel project.

I'll give a look at you ghidra files (I have a few radare2 projects I used back then for this disassembling nonsense :-) )

Still need to document this, some day.

Thanks.
David

[rplabs]

Hello,

Sorry to post in this old thread but I think it is better to have all info in one thread.

I have an EIP 578B later version it seems. Date code on the ICs are from 1994.
I has the options 05 and 06. I wanted to add option 02 and I followed the info here but is not working as it should. I added the AD7524 and the CR3 diode and removed R39 and R40 but I still had Error 13. So I started to investigate the CAL EEPROM X2816CP.

Offset 0x0 had 0x55 so I modified it to 0x00. Now option 02 is active and it doesn't give the Error 13 but is showing meaningless digits.
For example at 0.0dBm from my generator I see 096 at -30dBm I see 362 at +13dBm I see 268.

It never shows the dot or the minus sign. I guess I need some calibration values in the EEPROM but I cannot find any CAL EEPROM from a 578B.

I also dumped my firmware from the other EEPROM 27C512 and it is dated 1994.

I attached both files here in case someone needs them

Thanks!