| Electronics > Repair |
| Understanding and Repairing Clock Generator on Adret 742A UHF Generator |
| << < (5/10) > >> |
| PerArdua:
Thank you for the clarification timeandfrequency. Given that the CPU PCB has been cleaned up, and fresh glue logic installed, I feel slightly more confident with returning into the instrument. It is clear that the whole instrument is very interlinked. As mentioned previously, maybe the CPU is waiting for some other signal before continuing. As such I decided it was perhaps best to ensure the power supply was functioning correctly. Reading the schematic (I attached a couple I have put together and resized to fit a single image), I realise the "inhibit (CPU schematic)/Enable (PSU schematic)" should actually be shorted to the floating GND - annoyingly called -5V on the PSU schematic - to turn on non standby voltages. As below: Inhibit/Enable HIGH/floating: +15V pilot ---- ON (makes sense to keep the OCXO warm) +15V --------- OFF -15V ---------- OFF +5V ----------- OFF floating +5V -- OFF Dig_P +5V ---- ON (referred to as 5P on CPU schematic) Shorting inhibit/enable to floating GND (marked as -5V on the PSU schematic) turns everything ON. It's an interesting PSU schematic and nice to refresh a bit of knowledge on differential amplifiers, however it's apparent to me that Adret perhaps didn't have the most cohesive nomenclature/artstyle guidelines as there are numerous differences between different schematics that make understanding the system more difficult. 3 years does separate the drawings though... :) I don't believe Adret included the part numbers for the PSU motherboard components. These are (noting they use Qx and SNx multiple times on the same artwork) : Q1 - 2N3772 Q2 - 2N3772 SN1 - LM337K SN2 - LM350K All supplies are within +-200mV, I will look to carefully adjust the feedback potentiometers tomorrow to bring them in a bit tighter - though this is with no load. Thus far it's been left running with no load for 20 minutes and no funny business, therefor I'm confident in hooking this up with some new ribbon cable. I'll spend tomorrow working towards reading the ROMs. I thought I would have a go with using an Arduino to read them as the EPROM reader was a bit more expensive than I was anticipating. All the best, :-+ |
| George Edmonds:
Hi Per Ardua Sorry for the delay in replying to you, I had kidney surgery in December last year and am still recovering from it. In the past I have repaired a number of Adret 700 series instruments, both for normal failures and for leaking Ni-Cad batteries. I know that you are probably well aware that the material that leaks from the Ni-Cad is both highly corrosive and conductive. You can bet that although you did not power the instrument up the previous owner/s undoubtedly did causing allsorts of consequential damage. I have in the past successfully repaired one with about the same level of contamination of the CPU board that your photograph shows. Following the removal of the Ni-Cad battery I treated the board by brushing on to it white wine vinegar, the contamination is strongly alkaline so I used a mild acid to neutralise it, could also have used Citric acid, but I had none to hand at the time. I then washed the board multiple time with a strong surfactant and IPA, finally rinsing the board with deionised water. What I did find when I started to test the CPU board was that the contamination had not been removed from under the line of IC’s that go across the board, I had to replace them all using turned pin IC sockets. There was also some damaged track work that had to be repaired using kynar insulated wire. The real worry was that the contamination had managed to get under conformal coating and was starting to eat away the track work. I do know that this repair lasted for at least two years after which time the instrument was sold on , to whom I do not know and I lost track of it. As to the power supplies, the designer should have been sent back to do basic electronics 101 all over again. Originally the diode used were 5401, would have died like May flies, then 5404 which would also have failed due to over current from the inrush current to the large electrolytic capacitors they were feeding. I called the main power supply electrolytic capacitors, pink ones of unknown manufacture, used “magic capacitors” they used to go to zero capacitance due to no electrolyte in them , but where the electrolyte went was magic, there was no sign of it anywhere in the instrument, it just disappeared. To then use harmonica style connectors to link the PSU to the motherboard was the height of stupidity. Good plan if you want to constantly get burnt connectors, every 700 series instrument I have seen has had burnt power supply to motherboard connectors. Now you asked a question regarding EPROM’s and EEPROM’s. Now these comments are not Adret specific, but apply to most computer controlled signal generators. Normally the instruments operating firmware is held in EPROM’s up until say the 1990’s when flash memory started to be used. EEPROM’a or EAROM’s are normally used to hold the instruments calibration constants, passwords and instrument serial number etcetera which are unique to that one instrument. There are two problems with EEPROM’s, they have a maximum number of write cycles before they fail and by far more importantly they need write protection while the supply rails stabilise and from drooping supply rails The standard way of doing this is to monitor the +5V supply, if it is not correct the CPU reset pin is asserted to hold the CPU in a continuous state of reset thus preventing any false writes to the EEPROM. For the above reason alone I suspect that the CPU board will not operate outside of the main instrument. The use of decidedly none standard and unobtainable connectors for the CPU and other PCB’s is a real PITA. I was never able to find any to make a PCB extender. Hopefully someone may know a source of them. I do have a Marconi-Adret branded instrument which probably has a copy of the last firmware used, it is version 742.0013 for the EPROM’s and 742A-0015 for the EEPROM. There is also a second CPU board with EPROM’s version 7.1.88, the EEPROM has no version number, but is a 2816A. It has taken me some time to write this posting so other events may have taken over in the meantime. Hope the information helps and that you do finally get the signal generator working again. G Edmonds |
| PerArdua:
Thank you George, I hope you are recovering well from your surgery, thank you for taking the time and energy to help with my repair. I am, perhaps naively, optimistic about repairing this - even if I might need to see about getting a replacement PCB made up, it will all be an interesting learning experience. I am still (kind of) fresh out of university, trying to better my limited knowledge with real world examples. To that end, this repair has already been more useful than the more simple dried capacitor/broken op amp I've dealt with in previous repairs. In terms of mitigating the effects of inrush current on the diodes, what design differences could/should have been made? I can imagine using a NTC thermistor or a relay to switch in/out a low resistance series resistor was within the realms of possibility in the early 80s - were these prohibitively expensive during that time? Regarding the custom connectors, I've found that standard 0.1" pitch headers fit, providing you pull out/cut every other connection to permit mechanical clearance. This works for both male and female connectors, though perhaps there is some risk of overly large male headers damaging the female Adret connectors. I've drawn up some plans for an extender based on using these modified headers, if I need to purchase them later in the repair I'd be happy to send a spare PCB to you for free. |
| timeandfrequency:
Hi everybody, --- Quote from: George Edmonds on February 05, 2023, 05:56:51 am ---The use of decidedly none standard and unobtainable connectors for the CPU and other PCB’s is a real PITA. I was never able to find any to make a PCB extender. Hopefully someone may know a source of them. --- End quote --- I just had a quick look to the CPU board picture taken by PerArdua in an earlier post. IMHO, the CPU board simply uses a vanilla 31 pins connector which can be purchased off-the-shelf from many common vendors. These connectors were commonly used in (now vintage) TE. I even managed to dig up an extender and some additionnal connectors from my storage barn (just need to remove the two plastic brackets to make it usable) . And it seems rather easy to duplicate this part : drawing the artwork, manufacturing the PCB at PCBWay or JLCPCB, buying two connectors and a bunch of test pins to cobble such an extender should be doable for less than 20 bucks. [UPDATE #1] --- Quote from: PerArdua on February 05, 2023, 12:50:02 am ---Inhibit/Enable HIGH/floating: +15V pilot ---- ON (makes sense to keep the OCXO warm) +15V --------- OFF -15V ---------- OFF +5V ----------- OFF floating +5V -- OFF Dig_P +5V ---- ON (referred to as 5P on CPU schematic) Shorting inhibit/enable to floating GND (marked as -5V on the PSU schematic) turns everything ON. --- End quote --- Great! The test you made matches with the text explanation provided in then SM. I did not dig into the PS schematics yet. [UPDATE #2] --- Quote from: PerArdua on February 05, 2023, 01:56:16 pm ---In terms of mitigating the effects of inrush current on the diodes, what design differences could/should have been made? I can imagine using a NTC thermistor or a relay to switch in/out a low resistance series resistor was within the realms of possibility in the early 80s - were these prohibitively expensive during that time? --- End quote --- 1N5401 & 1N5404 have a maximum peak foward allowed surge currect of 200A, which is not that bad. Even I did not verify, I guess that Siemens (later Epcos, now TDK) already sold power NTC's and PTC's to mitigate surge currents in the early 80', but I would say that such kind of parts were not commonnly used in 50/60 Hz transformer PS. I've no idea if there where expensive at that time. Furthermore, it's not trivial to choose the best suited power NTC or PTC to fit inside a PS. It depends on the transformer & electrolytics caracterstics and also the shape of the inrush current you want to obtain (not too steep, not too flat). Sometimes, it's easier to buy a bunch of different values and give it a try... If you really want to dig into that topic, you may check out these links : https://www.ametherm.com/inrush-current/inrush-current-faq.html https://www.ametherm.com/inrush-current/ptc-thermistors-for-inrush-current-limiting https://www.ametherm.com/ptc-thermistors-vs-ntc-thermistors-for-inrush-current https://www.ametherm.com/inrush-current/calculators If you are concerned with the limited inrush current of the legacy parts inside the Adret, choosing diodes with better features can also be an acceptable solution. This TO220 package, 2x5A, 150V diode has a maximum allowed peak inrush current of 620A. Even if you recap your PS with first tier electrolytics, 620A won't be reached. |
| George Edmonds:
Hi I must disagree, we are talking about the CPU card to motherboard connector NOT the standard IDC connectors on the CPU board. The CPU card to motherboard connector I have never seen used elsewhere in my sixty years of experience. The parts list gives the manufacturer as Trelec, model TM31, When I last checked about three years ago this manufacturer was still in business, but no longer supplied the type of connectors used in the Adret 700 series instruments. It may well be possible to make up a suitable connector, BUT the pin and socket diameter are crucial if damage to the existing connects is to be avoided and good connections made. G Edmonds |
| Navigation |
| Message Index |
| Next page |
| Previous page |