willem programmer power modification.

Thread Starter

sam@2009

Joined Dec 21, 2023
30
Hi all.

I was thinking to make a big modification to one of my willem programmers , removing the power supply i mean the voltage booster and the 2 vcc diodes , and replace them with with 2 or 3 power regulators, the first regulator lm7805 will handle only ICs the second regulator lm317 gonna have 3 voltages 12.5/21/25 with trimmer, and the third regulator will have 3 voltages and gonna take care of vcc 4.5/5/5.6 also with trimmer . im not sure if it can work or not just im curious, and wanna want give a try.
i ve attached an picture of a power supply just to give an example.

any ideas will be welcome.
Thanks.
 

Attachments

dl324

Joined Mar 30, 2015
16,896
any ideas will be welcome.
I'd do what Willem did on the 4.0. He used LM317 for VPP and VCC. For VCC, he designed for 3 different voltages.

When I modify PCB3 variants, I give 2 fixed and 1 variable voltage for VPP and VCC.

I also replace the transistor used to switch VCC to the programming sockets with a relay. This addresses the problem with VCC dropping for "high" current devices (NMOS EPROMs, microcontrollers, higher capacity CMOS).

On the 4.1, he added a 3rd voltage regulator that supplied the logic. That made supporting low voltage devices more robust. On other programmers, they simply reduced VCC for the logic and device being programmed. Since the parallel port operates at 5V, going much below that can affect the interface to the computer.

EDIT: Willem's schematics weren't the neatest. Here's my redraw of the power section of the 4.1:
1704055923785.png
The 4.0 didn't have the LM317L or the relay.

Excerpt from Willem's 4.0 schematic:
1704056035264.png
 
Last edited:

dl324

Joined Mar 30, 2015
16,896
Picture of a modification I did for a PCB3 programmer. I removed the USB connector to mount the supply board.
1704056540539.png
 

Thread Starter

sam@2009

Joined Dec 21, 2023
30
hello everybody and thanks for your posts , to mister bill 2 these modifications will solve vpp , vcc voltage drop problems on this programmer .i have a wilem programmer with 4.75 vcc which means i t can never write a code on a eprom which needs vcc 6.20v i giving me a troubles also when it comes to reading it reads ok but it gives errors when veryfing.
 

dl324

Joined Mar 30, 2015
16,896
i do 27c322 27c160 never been able to program these devices i do also 2732.
Those are some of the "high" current devices I referred to.

27C322 -> 70mA
27C160 -> 70mA
2732 -> 125mA

Unmodified Willem programmers can't read or program these devices reliably.
 

dl324

Joined Mar 30, 2015
16,896
very interesting work dennis i will study the shematics ,and i will build it.
If you look at the area circled in red below, that's a 6 pin female header I installed on a Willem 4.0 to give more voltage options and replace the transistor used to switch power to the programming sockets with the relay mod:
willem4v0Mod.jpg
The wiring on perfboard was to modify the programmer to be compatible with the 0.98 program versions.
 

dl324

Joined Mar 30, 2015
16,896
Before you modify your programmer, you can do some tests to show how bad power is on the PCB3 variants.
  1. Go to the Test H/W tab.
  2. Set VPP to 12.75V (required by Intel 27512)
  3. Enable VPP, measure voltage, then disable.
  4. Install a 270 ohm resistor between pins 1 and 16 on the ZIF
  5. Enable VPP and measure voltage at pin 1. It should be 12.0V or higher. Desired current 50mA.
For 21V (Intel 2732A), use a 750 ohm resistor. Voltage shouldn't drop below 20.5V. Desired current 30mA.

For 25V (STM 2716), use a 820 ohm resistor. Voltage shouldn't drop below 24V. Desired current 30mA.


Do similar tests for VCC.

For 27512, set VCC to 6.25V, measure voltage on pin 32. Install a 42 ohm resistor from pin 32 to 16, voltage shouldn't drop below 5.75V. Desired current 150mA.

For 2732A, set VCC to 5V and measure. Install a 50 ohm resistor, voltage shouldn't drop below 4.75V. Desired current 100mA.

Current requirement for 2716 is the same as 2732A.


Voltage and current requirements in tabular form:
1704141356815.png

You could use 1/4W resistors, but some will be dissipating significant power, so do the measurements quickly.


If you want to determine which EPROMs you can read reliably, set VCC to 5V and decrease load resistance until voltage drops to 4.75V. Calculate the current the programmer could provide and compare that against any device you want to read. For example, M27C160 requires up to 50mA.


Another limitation in the programs. 2716/2732 require a nominal programming pulse of 50ms. The 0.97g can only do 12ms max. I compensate by setting VCC to 6.25V (but I wouldn't go any closer to the 7.0V absolute maximum).
 
Last edited:

dl324

Joined Mar 30, 2015
16,896
I tested my Dual Powered PCB3:
1704319894716.png
Based on STM devices, 1995 databook.

My programmer has a molded inductor (which saturates at too low of a current). The flyback diode doesn't look like a Schottky diode. A PCB3 schematic I have shows a 1N4148, which won't work well.

The schematic also shows a rectifier diode on the output of the 5V regulator; which is a terrible idea:
1704320311689.png
The diodes in the ground path of the 7805 are also a bad idea. That regulator wasn't designed to have a stable current from the ground pin.

On my programmer, LED2 is green. The schematic shows programming VCC green when most programmers would use yellow (at least that's what Willem and the guy who designed ATH3.1 did).
 
Last edited:

dl324

Joined Mar 30, 2015
16,896
I tested my Willem 4.0 with the power modification I gave you.
1704337247232.png

I'm surprised that there was such a large voltage drop at VCC=6.25V. Relay contact resistance is supposed to be 0.1Ω max and I calculated the total resistance to be almost 0.4Ω, so there's about 0.3Ω of wiring resistance.

It's still much better than any PCB3 variant. They can't supply anywhere near 180mA.

There are some Willem type programmers that can have VCC=6.4V with 2 diodes in series with ground for the 5V regulator (but no device specifies 6.4V for anything) and some that can go as low as 4.3V (or lower). STM specifies a verification pass at 4.2V.
 
Top