Topics

2445 EPROMs

christopherbath@...
 

I need to replace the EPROMs in a 2445 oscilloscope. These appear to be two 27128 EPROMs with part numbers 160-1994-09 & 160-1995-09. Does anyone know the access time required for these? I am able to purchase 200ns access time 27128 but wanted to be sure these are sufficient. The existing EPROM part number is covered and I wanted to keep this intact in case I need to reuse these.

Thanks,

Chris

Panos
 

christopherbath@...
 

Hi,

Thankf for this. I still would like to find out the access time for the 27128 EPROMs that are used in the instrument. Does anyone know what the this is?

Thanks,

Chris

Mark Litwack
 

Hi Chris,

The 2445 and 2465 have Intel D27128 (no suffix) EEPROMs. This is a 250ns part.

Gentle warming with a heat gun can help peeling up all or part of a label without damaging it.

-mark

Chuck Harris
 

The processor in the 2445 is a 2MHz 6802 microprocessor.
It needs 250ns or faster memory.

Any 27128 you find is fast enough.

Why do you think you need to replace them?

-Chuck Harris

christopherbath@... wrote:

I need to replace the EPROMs in a 2445 oscilloscope. These appear to be two 27128 EPROMs with part numbers 160-1994-09 & 160-1995-09. Does anyone know the access time required for these? I am able to purchase 200ns access time 27128 but wanted to be sure these are sufficient. The existing EPROM part number is covered and I wanted to keep this intact in case I need to reuse these.

Thanks,

Chris



christopherbath@...
 

Hi Chuck,

Looking at the schematics the processor seems to be driven by a 5MHz clock (the clock being generated by a 10MHz crystal). According to the datasheet for the 6802 the maximum clock frequency is 4MHz so I am not sure what is going on here? Possibly the processor is a MC68A02 or MC68B02. I guess you would need to look into the bus timing is more detail to work out the access time requirements for the ROM but agree that 250ns or less would be more than enough.

The 2445 won’t power up properly. It seems to me as if the processor hangs on start-up. Given that the scope seems to do slightly different things on start-up I am suspicious about the EPROMs that would have been programmed about 30 years or more ago. The power supply rails all seem to be ok so I think the EPROMs are the next place to be looking. Perhaps I should also check the clock and reset for the microprocessor.

Regards,

Chris

Chuck Harris
 

Hi Chris,

My humble apologies, the CPU uses a 5MHz square wave clock. The
processor is a 68A02.

Early microprocessors usually took several cycles per instruction.
5, I think was the minimum for the 6800 family... leading to a 200ns
requirement on access time.

It has been a long, long, time since I last had to design with, or
debug a 6800 series microprocessor. I have never found one bad in
all of the 2465's I have calibrated... actually, I have never found
one bad in anything I have worked on.

200ns is what tektronix used as the access time for the RAM, and EPROM.

I would put the probably that the EPROM is good at 99.999%... unless
you have had a power supply problem that has elevated the 5V to above
7V... in which case, most everything on the CPU card will be bad.

Look instead at the jumpers on the board. There are test jumpers that
will prevent boot up.

Also, look at the reset circuitry. A single bad part, and the CPU
will never get out of power on reset. Pay particular attention to the
0.1uf and 2.2M resistors that provide the timing.

Assuming that there is CPU activity, there should be an LED lit on the
front panel indicating the failure cause.

-Chuck Harris

christopherbath@... wrote:

Hi Chuck,

Looking at the schematics the processor seems to be driven by a 5MHz clock (the clock being generated by a 10MHz crystal). According to the datasheet for the 6802 the maximum clock frequency is 4MHz so I am not sure what is going on here? Possibly the processor is a MC68A02 or MC68B02. I guess you would need to look into the bus timing is more detail to work out the access time requirements for the ROM but agree that 250ns or less would be more than enough.

The 2445 won’t power up properly. It seems to me as if the processor hangs on start-up. Given that the scope seems to do slightly different things on start-up I am suspicious about the EPROMs that would have been programmed about 30 years or more ago. The power supply rails all seem to be ok so I think the EPROMs are the next place to be looking. Perhaps I should also check the clock and reset for the microprocessor.

Regards,

Chris



Mark Litwack
 

Hi Chuck and Chris,

The external clock input on the 6802 is divided internally by four to create an internal two-phase clock (one phase is output as signal "E"). So, with a 5MHz input, the cycle time is actually 800ns, which well within range of the Intel D27128 with 250ns access time I mentioned before.

There is also the NOP exerciser which can be enabled with jumper P503. This is good for getting to the bottom of possible processor and/or bus buffer issues by looking at E, address lines, and other signals as the processor is looping. Its operation is described in the service manual.

I do recall one instance of a bad processor, but granted, they are rare.

-mark

Chuck Harris
 

Back in those days I was always and Intel guy.

Customers forced me to learn Motorola 6800 and 68000, but
it wasn't a labor of love, so I have forgotten almost
everything I ever knew about them.

And my data books for them are ...somewhere...

Thanks Mark!

Mark Litwack wrote:

Hi Chuck and Chris,

The external clock input on the 6802 is divided internally by four to create an internal two-phase clock (one phase is output as signal "E"). So, with a 5MHz input, the cycle time is actually 800ns, which well within range of the Intel D27128 with 250ns access time I mentioned before.

There is also the NOP exerciser which can be enabled with jumper P503. This is good for getting to the bottom of possible processor and/or bus buffer issues by looking at E, address lines, and other signals as the processor is looping. Its operation is described in the service manual.

I do recall one instance of a bad processor, but granted, they are rare.

-mark



Harvey White
 

that no-op sequence is used for signature analysis.

Harvey

On 1/17/2020 9:47 AM, Mark Litwack wrote:
Hi Chuck and Chris,

The external clock input on the 6802 is divided internally by four to create an internal two-phase clock (one phase is output as signal "E"). So, with a 5MHz input, the cycle time is actually 800ns, which well within range of the Intel D27128 with 250ns access time I mentioned before.

There is also the NOP exerciser which can be enabled with jumper P503. This is good for getting to the bottom of possible processor and/or bus buffer issues by looking at E, address lines, and other signals as the processor is looping. Its operation is described in the service manual.

I do recall one instance of a bad processor, but granted, they are rare.

-mark


Mark Litwack
 

NOP + signature analysis was prevalent for servicing HP gear from that era, but I'm not aware of any signature procedures created for the 2445/2465. (If there is, please share!)

Even without signature analysis, the NOP exerciser in any equipment is a great mode for examining proper functioning of the address lines, decode logic, and other signals related to the processor. All you need is a scope.

-mark

On Sat, Jan 18, 2020 at 12:34 AM, Harvey White wrote:


that no-op sequence is used for signature analysis.

Harvey


On 1/17/2020 9:47 AM, Mark Litwack wrote:
Hi Chuck and Chris,

The external clock input on the 6802 is divided internally by four to create
an internal two-phase clock (one phase is output as signal "E"). So, with a
5MHz input, the cycle time is actually 800ns, which well within range of the
Intel D27128 with 250ns access time I mentioned before.

There is also the NOP exerciser which can be enabled with jumper P503. This
is good for getting to the bottom of possible processor and/or bus buffer
issues by looking at E, address lines, and other signals as the processor is
looping. Its operation is described in the service manual.

I do recall one instance of a bad processor, but granted, they are rare.

-mark



Harvey White
 

And I don't know all that much about the 2445/2465.  So I have no idea if they ever used signature analysis in those scopes. However, throughout the TM500 (with CPU) and the TM5000 series, I have seen signature analysis numbers, at least on the ones I have.  Haven't gone too deeply into some of the HP gear, so even though they had a signature analysis tool (5005B for me), can't say which HP equipment there is.

Some work can be done to get an idea of what the signature might be on unknown nodes, but I'm not sure how to do that.  Never needed to do so.

Harvey

On 1/18/2020 4:05 PM, Mark Litwack wrote:
NOP + signature analysis was prevalent for servicing HP gear from that era, but I'm not aware of any signature procedures created for the 2445/2465. (If there is, please share!)

Even without signature analysis, the NOP exerciser in any equipment is a great mode for examining proper functioning of the address lines, decode logic, and other signals related to the processor. All you need is a scope.

-mark

On Sat, Jan 18, 2020 at 12:34 AM, Harvey White wrote:

that no-op sequence is used for signature analysis.

Harvey


On 1/17/2020 9:47 AM, Mark Litwack wrote:
Hi Chuck and Chris,

The external clock input on the 6802 is divided internally by four to create
an internal two-phase clock (one phase is output as signal "E"). So, with a
5MHz input, the cycle time is actually 800ns, which well within range of the
Intel D27128 with 250ns access time I mentioned before.
There is also the NOP exerciser which can be enabled with jumper P503. This
is good for getting to the bottom of possible processor and/or bus buffer
issues by looking at E, address lines, and other signals as the processor is
looping. Its operation is described in the service manual.
I do recall one instance of a bad processor, but granted, they are rare.

-mark



christopherbath@...
 

Hi Mark & Others,

I have finally managed to program new EPROMs after a bit of a nightmare in having misplaced my old programmer and also loosing the software for another one.

The bad news is the re-programmed EPROMs don't fix the issue. I was starting to get a bit suspicious now about the processor itself which is a S68A08 and having looked around on the web for a possible replacement I found a few other posts where people had issues with the S68A08 processor in the 2445.

The IRQ waveform is not as expected as per the trouble shooting procedure in the service manual. I have run the NOP and the address lines do seem to have a binary count on them but I am still not convinced yet that the processor is working correctly.

Do you have any recollection of what the faulty processor was you had mentioned previously? This one is an AMI (American Microsystems) S68A08. From what I can tell that specific microprocessor was not very common.

Sorry for the very belated reply.

 

On Sat, May 2, 2020 at 02:32 PM, <christopherbath@...> wrote:


Do you have any recollection of what the faulty processor was you had
mentioned previously? This one is an AMI (American Microsystems) S68A08. From
what I can tell that specific microprocessor was not very common.
AFAIK AMI's S68A08 was just a second source for the Motorola MC68A08, IOW an MC68A02 with unusable/defective/disabled internal (128 byte) RAM. The A refers to a version with a slightly higher maximum clock frequency.
I'm not sure if a 68A02 (or 68B02) would just work but that should be easy to find out, because the data sheets are available on the 'net.

Raymond

Chuck Harris
 

It's not like it is hard to get an A5 card for a
2465. PM me an address, and I will send you one
for the postage.

-Chuck Harris

christopherbath@... wrote:

Hi Mark & Others,

I have finally managed to program new EPROMs after a bit of a nightmare in having misplaced my old programmer and also loosing the software for another one.

The bad news is the re-programmed EPROMs don't fix the issue. I was starting to get a bit suspicious now about the processor itself which is a S68A08 and having looked around on the web for a possible replacement I found a few other posts where people had issues with the S68A08 processor in the 2445.

The IRQ waveform is not as expected as per the trouble shooting procedure in the service manual. I have run the NOP and the address lines do seem to have a binary count on them but I am still not convinced yet that the processor is working correctly.

Do you have any recollection of what the faulty processor was you had mentioned previously? This one is an AMI (American Microsystems) S68A08. From what I can tell that specific microprocessor was not very common.

Sorry for the very belated reply.