Willem 5.0 Eprom Programmer

Discussion in 'General Electronics Chat' started by payty, Jun 24, 2015.

  1. payty

    Thread Starter New Member

    Jun 24, 2015
    12
    0
    Hi,

    I have this Willem 5.0 Eprom programmer that I have bought from Ebay.
    I have bought it for some older eproms that I need to program.

    With newer 27C256 and 27C512 it works ok, but with older 27128 and 2716 it does not because of the voltage drop on VPP.

    It has a potentiometer (in the red square) that can be used for VPP adjusting but under load I am unable to do that. It does not go higher than ~10V.

    I have tried both USB power and external power (from a laptop PSU that can provide max 5A).

    The inductor for the step-up part is in the green circle in the picture. It looks like a small capacitor,nothing written on it.


    @dl324 -> thanks for the info in the other thread.
    I have found this site :
    http://www.py2bbs.qsl.br/willem_pcb45en.php
    where the author describes an inductor that should replace the original (number of turns, diameter, etc). But when I have replaced mine with that one, max VPP is even lower, so it seems that it is not ok.
    This is the picture of the willem. Any advice?
    willem.png
     
  2. dl324

    Distinguished Member

    Mar 30, 2015
    3,234
    619
    It looks like your programmer is using a better inductor than most PCB3 variants and it may be as good as the ones you referenced on eBay. Can you read the values for the inductor and timing cap? The timing cap is connected to pin 3 of the MC34063A. When I was experimenting with the step up regulator on one of my PCB3 variants, I used a J.W.Miller 5900-181 because I happened to have some in my parts bin. That inductor can handle 1.6A, with a saturation current of 2A.

    You should stop using the USB power option. It can only be used for reading EPROMs. For programming anything larger than 32Kb, you require VCC > 5V. For 16Kb and 32Kb devices, you can program with VCC=5V, but your programmer will have a VCC voltage drop problem on NMOS EPROMs (and higher capacity, like 8Mb, CMOS devices).

    It looks like your programmer has several VCC options. The two yellow jumpers by the ZIF socket are probably for sub 5V VCC and the green jumper near the 7805 may give you a couple of VCC options higher than 5V. You need the sub 5V options for proper blank verification and you need the higher options for programming.

    You're still going to have a VCC drop problem with NMOS EPROMs. Some have tried to replace the VCC transistor switch with one with a lower saturation voltage. Willem and I decided a relay was the best solution. At 100mA, you get around a 10mV drop vs. 0.5-2V using a transistor.

    Thanks for reminding me about that website. He has a lot of useful information on Willem type programmers.
     
    Last edited: Jun 24, 2015
  3. payty

    Thread Starter New Member

    Jun 24, 2015
    12
    0
    Hi,

    On the inductor I could not read anything (nothing written).
    The capacitor is 100nF.

    I have some electronics knowledge but not much. On the PCB3.0 schematic, the inductor is 100uH.
    Should I stick to this value of the inductance or higher is better? When I have built that coil mentioned on the site above, I have realized that when I have added more turns to the specified 55, the max current was even lower.
    For me it is easier to build the coil as I can experiment with it but it seems not to be working. What should I change? Number of turns or toroid size?

    Should I change other components from the step up regulator according to the inductor I use?

    On ebay I can find several inductors with different inductance but not many other characteristics. They only give inductance and rated current (I assume that it is different from saturation current).

    But for others I see that they give more data:
    http://www.ebay.com/itm/5pcs-100uH-...530?pt=LH_DefaultDomain_0&hash=item2352f6584a


    They say 100uH, rating current 2.5A, saturation current 4A.



    Or these
    http://www.ebay.com/itm/Lot-de-3-in...74?pt=LH_DefaultDomain_71&hash=item2801f85136


    Best regards,
     
  4. Hypatia's Protege

    Distinguished Member

    Mar 1, 2015
    2,778
    1,218
    Be advised also that many older 27 series devices are specified at a 21V VPP...

    Best regards
    HP
     
  5. dl324

    Distinguished Member

    Mar 30, 2015
    3,234
    619
    A little larger inductance would be better than smaller. What inductance did you expect with whatever turns you put on your toroid?

    With no load, you should have been able to get 21V; 25V is sometimes iffy...

    It looks like the diode on the regulator is a 1N4148 or equiv. I've found that diode to be okay, but some have replaced it with a faster, possibly lower forward voltage, diode.

    The 100nF timing cap is probably too large; 200pF to 1000pF is more typical. Are you certain you got the right cap? The decoupling caps are usually 100nF.
     
  6. payty

    Thread Starter New Member

    Jun 24, 2015
    12
    0
    I do not know about the inductance. I just took the figures from that site, built the inductor and saw that it was not better than the original.

    And then I have rebuilt it with more turns but it was worse than before.

    And in all cases I could not trim the VPP under LOAD over a threshold ->with original inductor ~10.5V, second inductor ~8.5(55 turns of 0.5mm wire) and the last ~7 (with 90 turns of 0.5mm wire)


    I will look again after the capacitor when I get back home from my office. That one that I have mentioned is connected to PIN3 but maybe there is another one.

    Maybe changing that to a lower value could help.
     
  7. dl324

    Distinguished Member

    Mar 30, 2015
    3,234
    619
    What is the inductance index of your core, core type, and how many turns did you try?
     
    Last edited: Jun 25, 2015
  8. payty

    Thread Starter New Member

    Jun 24, 2015
    12
    0
    About the inductance index, I do not know. I took a core that I have found in my boxes with other electronic parts. I think it is ferrite, but coated with a white paint.
    Dimensions are:

    Height - 5mm
    ID-8mm
    OD- 14mm

    With 55 turns -> max 8.5V
    with 90 turns -> ~7V
    with 40 turns it seems to go better ~9.5

    Meanwhile I have checked the MC34063 datasheet and it seems very ok, with a lot of info.
    There are some example schematics of step-up and step-down converters.
    That step up can deliver 175mA with a 170uH inductor. The datasheet contains also data about the core and number of turns for the inductor.

    I will try to modify the Willem according to this one.
     
  9. dl324

    Distinguished Member

    Mar 30, 2015
    3,234
    619
    Can you post a picture of how you were winding the inductor? Without knowing the value of the inductor, you'll be shooting in the dark...

    Once you get the step up regulator working correctly, you're still going to have a severe VCC voltage drop problem with 2716 and 27256. You will not be able to program them reliably; even if the program indicates otherwise. You cannot verify program margin without a stable, known VCC.
     
  10. payty

    Thread Starter New Member

    Jun 24, 2015
    12
    0
    willem.png Well, after doing some more digging, it turns that the device is not so bad.

    It's an improved version that should take care of all the voltage problems. I have found that there are some jumpers on the PBC for this.

    In the picture, the yellow square jumpers are used for modifying the VPP. So With the trimmer I can fine-tune it, at 12.5V. And by changing the position of the jumpers, the VPP becomes 21.5 or 25V. And it stays there, there is no VPP drop. I have been able to write yesterday some old 2716 (KR573RF5 is the code) Soviet Union chips at 21.5V.

    Only one I could not write, a 27128, no matter what I have tried. But I think that one is broken.


    And for VCC, there is the jumper with the pink square. And I can provide three VCCs from there (5V, 5.5 and 6.2)
     
  11. dl324

    Distinguished Member

    Mar 30, 2015
    3,234
    619
    It never occurred to me to ask you if you were setting the jumpers correctly. All PCB3 variants are based on the same step up regulator schematic, so they use 2 jumpers to set 4 VPP voltage options...
    It's almost certain that you'll have VCC drop problems. Individuals working on Willem designs didn't know much about analog design and typically chose a BC557 PNP transistor that isn't appropriate for switching VCC, and is marginal for VPP.

    The worst case current for NMOS EPROMs is 100mA from VCC and 50mA from VPP. To check for the VCC drop problem, you can enter the test hardware screen (be sure to set the jumpers as indicated on that screen) and installing a 100 ohm resistor (1/2W or larger) between pin32 and pin16. Set VCC to 5V, enable VCC by clicking on pin32 on the screen, and measure the voltage. At 100mA, which happens to be the max collector current for BC557, saturation voltage will be a couple volts. The image below is from the datasheet:
    bc557VceSat.jpg
    The saturation voltage is > 1V at 35mA... Some have tried to compensate by increasing VCC. But that's a hack because proper program margin checking requires a known, stable voltage; not one that varies non-linearly with current.

    Typical VCC spec for programming 2716 is 5V +/- 5%, so anything below 4.75V is unacceptable. Using 5.25V is better because extra program margin will reduce problems with random bit flips in electrically noisy environments.

    You have a similar problem with VPP voltage drop, but it isn't as critical as long as you check program margin on the devices you program. If you don't know what program and erase margin are, you probably shouldn't be programming EPROMs for any critical applications.
     
    Last edited: Jun 26, 2015
  12. dl324

    Distinguished Member

    Mar 30, 2015
    3,234
    619
    What is the full part number for the devices you're programming?

    2716 require VPP=25V, 2716B require 12.5V.

    27128 require VPP=21V, 27128A require 12.5V.

    Since VPP is the highest voltage supported by the device, exceeding max VPP is very likely to damage the device...
     
  13. payty

    Thread Starter New Member

    Jun 24, 2015
    12
    0
    Today I have programmed some old Soviet Union 2716 equivalents. The part number is K573RF5. I could not find the datasheet, but they seem to be programmable at 21.5V.

    After programming and verification, the software says that verification is ok. My question is: should I trust this? If it says "OK" then I should expect that the chip contains the data I have been trying to write?
     
    Last edited: Jun 26, 2015
  14. dl324

    Distinguished Member

    Mar 30, 2015
    3,234
    619
    A VPP=25V part may program at a lower voltage. Whether they programmed reliably is a different concern.
    It depends. Do the voltage tests I suggested. If you don't have a voltage drop that causes you to drop below the programming specs, you used the correct timing pulse (50mS for devices smaller than 32Kb), and the program said verification passed; you're most of the way there. To check program margin, I'd set VCC to at least 5.5V and do a verify against the data you programmed. The absolute max VCC for those 2716 is probably 7V, same as for 64Kb devices that program at VCC=6-6.25V, so you could potentially verify margin to 6.25V.
     
  15. payty

    Thread Starter New Member

    Jun 24, 2015
    12
    0
    Hi,

    I have checked the voltages directly with some chips installed, using the HW Check window.

    Newer chips I have are at 4.98Vcc.
    The 2716 drop to 4.9
    Only for one NMOS chip, VCC drops to 4.8V - 4.82V


    With no load, measured VCC is 5.01V

    I have added a radiator to the 7805 regulator because during programming it gets hot.
     
  16. dl324

    Distinguished Member

    Mar 30, 2015
    3,234
    619
    If you're measuring with devices in the socket, that isn't a valid measurement. You need to measure at the worst case load to ensure it will work for all devices. Presumably, manufacturers would reject any device that exceeded the maximum currents in the specs.

    I did load measurements on one of my PCB3 variants and a modified Willem 4.0. With a 100mA load on VCC; the PCB3 dropped 0.72V and the 4.0 dropped 0.09V (it has the relay mod).

    What is the output voltage of the external power source?
     
  17. payty

    Thread Starter New Member

    Jun 24, 2015
    12
    0
    It's a universal laptop power supply that can provide between 12V and 24V. 90W power.
    For the willem I am using the 12V setting.

    Ok, then I need to measure with a resistor. I'll come back with the voltages soon.
     
  18. dl324

    Distinguished Member

    Mar 30, 2015
    3,234
    619
    The VCC drop on the PCB3 was lower than I remembered from previous experiments, so I checked some NMOS EPROM datasheets. I found that for 256Kb, max current from VCC was spec'ed at 150mA. For some 32Kb, it's also 150mA (Hitachi 2532), but for some 128Kb (ST Microelectronics), it's 100mA.

    With a 150mA load, VCC dropped 1.40V (to 3.67V) on the PCB3 and 0.12V (to 4.93V) on the modified 4.0. I also observed that the voltage drop on the PCB3 increased 0.01V every couple seconds before I removed the load, while it was stable with a relay. For reliable programming, the voltages need to be stable throughout programming and verification. For 2716, that would be a minimum of a 2 minutes.

    The bottom line is that the programmer needs to handle 150mA without a significant voltage drop. Otherwise, you could have reliability problems with devices with power dissipation at or near the spec'ed maximum.

    I also just noticed that I said a 100 ohm load should be put on VCC; that should have been 50 ohms and should be updated to be 33 ohms.
     
    Last edited: Jun 28, 2015
  19. dl324

    Distinguished Member

    Mar 30, 2015
    3,234
    619
    In TO-220, the 7805 can dissipate about 2W without a heatsink at an ambient temperature of 25C, so you're probably okay. When I modify a Willem programmer, I usually put a heatsink on the regulators just to be safe; you never know when someone will do something silly...
     
  20. payty

    Thread Starter New Member

    Jun 24, 2015
    12
    0
    Hi,

    With a 30 ohm resistor I get 4.8V for the VCC.

    Jumpers are set for 5V.
     
Loading...