test the IC 2519

I did a Google search and found this information:

Testing a shift register IC with a multimeter primarily involves checking for short circuits and verifying supply voltage (VCC) and ground (GND) connections. Set your multimeter to diode or continuity mode to check for shorts between VCC and GND. A near-zero reading indicates a likely dead, shorted IC.

Steps to Test a Shift Register with a Multimeter:

1. Identify Pins: Locate the VCC (positive) and GND pins using the datasheet. Generally, VCC is near a bypass capacitor.2. Power Off - Continuity Check: Turn off the circuit. Set the multimeter to continuity (beep) mode. 3. Check Short Circuits: Place probes between VCC and GND pins. A constant beep or near zero ohms resistance means the  IC is likely shorted internally.4. Diode Test (Power Off): Set to Diode Mode. Place the black probe on GND and red probe on input/output pins. A healthy  IC usually shows a voltage drop between 0.4V and 0.8V (400 - 800 mV).5. Power On - Voltage Test: Apply power. Set the multimeter to DC Voltage. Check that VCC pin is receiving the correct  voltage relative to GND.

A multimeter cannot fully test the logic function of a shift register (shifting bits). For that, you need to monitor output pins while providing a clock signal or use a logic analyzer, or connect the outputs to LEDs to see the data shifting.

Possible helpful links:https://www.applefritter.com/files/signetics2519.pdfhttps://www.utsource.net/itm/p/1884790.html

I hope this is helpful in some way. I am not knowledgeable on the 2519B IC. But understanding how the IC works is a good place to start (in my opinion). 

Joe

It has been my experience that a lot (40-50% ?)  of the 2519N I got in the early phase of my kits were bad, and about half of the bad ones were dead right out of the tube, and another half failed during the first ~3 days (or so) of the burn-in (running my diagnostics page in an Apple-1). Typically, these still showed something going on (on the screen) so they did not die, but limped along in some dysfunctional and useless way.

The takeaway from this is that if you buy 2519N, and there is only one source left (in Israel), that you get a 100% function / money back guarantee if it does not work fine in your Apple-1. And you need to do a long test at full clock speed and full output load, so these ICs run at the higher temperature and stress as they encounter in the Apple-1.

All my later kits used 2519B from another source (also in Israel, but a different seller) and these had very few bad ones (~ 3%) and all had the desirable 1976 date codes.

Buyer beware !

- Uncle Bernie

I supplied the -12V with an adapter and wrote a test-routine for the ic-test of my topmax programmer.

The 2519s require only a single clock signal, as an internal clock generator produces a three-phase clock.

The inputs are all TTL-compatible.

The outputs are either at a high level or have a high impedance – effectively an open-collector configuration with VCC!

You could, of course, pull them low with a resistor to ground – but the Topmax also recognises high impedance.

I plugged the shift register into a socket and removed pin 5. There, I fed in -17V externally against VCC.

In a test programme, I wrote ‘low’ 40 times to all 6 inputs (recirculate = 0) to clear all cells. Whilst doing so, I initially ignored the status of the outputs.

Then I wrote ‘High’ once and recirculated 40 times (input recirculate = 1). I checked that 39 times it showed high impedance and on the 40th time it output ‘High’.

Output 4 is then not ‘High’, but shows high impedance (photo). I tested this with both ICs – the result is the same for both.

I swapped outputs 4 and 5 using an adapter connected to the Topmax, and the error persists on pin 9 – so it’s down to the Topmax input, and output 5 is OK. It was exactly the same with my second Topmax unit – it’s probably a system-wide issue.

If I write a ‘Low’ instead of a ‘High’, the output correctly returns as high-impedance after 40 cycles.

In the photo, the last test line (80) shows the error. The test line is shown at the top – the 6 ‘H’s are the expected outputs, with the measured outputs below them. The ‘Z’ is Out4.

test2519_002.jpg

   

test2519_001.jpg

In post #6, 'vossi' wrote:

" You could, of course, pull them low with a resistor to ground – but the Topmax also recognises high impedance. "

Uncle Bernie comments:

I would advise that you are more careful with testing 2519. These are easily zapped. And exactly how the Topmax tests for high impedance ("tristate Z") also needs to be scrutinized. I'd recommend a test fixture which has the pulldown resistors on the outputs as prescribed in the 2519 datasheet - it's in the Signetics PMOS/NMOS databook of 1972, IIRC the year correctly. To avoid damage to the Topmax you may need to add Schottky clamping diodes on the outputs which have a forward voltage (Vf) of ~0.4V or less and this would limit the negative excursion on these outputs to the same voltage (Vf). Depends on the actual pin driver / pin sense circuits on the Topmax which I don't have, so can't say if this is safe enough for it.

Myself, I would not try to use a device programmer of any kind for such tests on ancient ICs that need "strange" power supply voltages other than 5V. Been there, done that, and then had to repair the pin drivers of the programmer. And there is another risk ... these programmers typically cycle the supply voltage (the 5V in this case) and if other supply voltages are supplied externally (beyond the control of the programmer), this may damage either the DUT, or the programmer, or both.

Readers should take this as a well meant warning. "Testing" ICs or other devices with improvised rigs has caused more damage that you would think. The most embarrassing event of this kind I have heard of involved "incoming inspection" of a prototype missile guidance system at a major defense contractor. The incoming inspector applied line voltages and destroyed it. Inspection done ! Unit found to to be defective (it's giving smoke signals). Back to the supplier it went  - "Rejected due to not passing incoming inspection." The supplier was not amused and sent their own technician with the replacement unit to prevent further sabotage of this kind. And in my own field (IC design), I saw enough mishaps with not fully debugged "load boards" on IC testers which in some cases destroyed all of the prototype ICs we had. "Testing" can be a killer ... it's safer to build an Apple-1 and just plug it in. After you have checked that all supply voltages and the clocks are in spec. --- very imporant for dynamic logic, never try to run it without proper clocks. For instance, if  in  a two phase dynamic logic system one clock "hangs" on the active level, and the other clock runs normally, half of the time both clocks are active and this may cause shoot-through currents in the logic which can quickly overheat and/or destroy the IC. I once had counterfeit DS0025 from China - which actually were re-labeled industry standard MOSFET gate drivers made in CMOS - and these turned on both the PHI3 and PHI4 clocks on the Apple-1  all the time, which cause the 2504 shift registers to overheat and die. The irony was that the CMOS IC had no pulldown resistor on its input (the real DS0025 made in a bipolar process has them) and the randomly picked up charges on these pins made it work normally for a short time, until the inputs had floated to an operating point where both drivers would turn the clocks on (-12V). And then the destructive events unfolded ...  

- Uncle Bernie