7 Nov
2022
7 Nov
'22
11:51 p.m.
I have three DEC RA8x drives that have failed (all of them fault with
"spin error") because of bad photo-interrupter tachometer sensors. After
talking to a few friends, it sounds like this is a pretty common fault.
Photo sensors like this are fairly common, even today, but the specific
parts DEC used are weird and unobtainable.
I designed a little PC board that uses an ITR9606 photo interrupter, a
2N3904 and a resistor as a replacement. Works great - gives a beautiful 5V
P-P clean waveform and with the PCB it's a mechanical drop in replacement.
Just screws right onto the original mounting holes and plugs into the
original connector.
I put the PCB design up on OSH Park
https://oshpark.com/shared_projects/z8DSkQsP
If anybody needs one you're welcome to order some PCBs and build your own.
The ITR9606 is as common as dirt, and you can buy bags of ten on Amazon for
a few bucks.
I've also put a few pictures on the Facebook DEC Computer Users group. I'd
post the pictures here, but I don't think cctalk accepts attachments. It's
not really worth making a web page for it.
Bob
8 Nov
8 Nov
12:01 a.m.
May I suggest that you post this to the DEC forum on the VCFED.org website?
On 11/7/2022 5:51 PM, Robert Armstrong via cctalk wrote:
I have three DEC RA8x drives that have failed (all
of them fault with
"spin error") because of bad photo-interrupter tachometer sensors. After
talking to a few friends, it sounds like this is a pretty common fault.
Photo sensors like this are fairly common, even today, but the specific
parts DEC used are weird and unobtainable.
I designed a little PC board that uses an ITR9606 photo interrupter, a
2N3904 and a resistor as a replacement. Works great - gives a beautiful 5V
P-P clean waveform and with the PCB it's a mechanical drop in replacement.
Just screws right onto the original mounting holes and plugs into the
original connector.
I put the PCB design up on OSH Park
https://oshpark.com/shared_projects/z8DSkQsP
If anybody needs one you're welcome to order some PCBs and build your own.
The ITR9606 is as common as dirt, and you can buy bags of ten on Amazon for
a few bucks.
I've also put a few pictures on the Facebook DEC Computer Users group. I'd
post the pictures here, but I don't think cctalk accepts attachments. It's
not really worth making a web page for it.
Bob
12:56 a.m.
On Nov 7, 2022, at 6:51 PM, Robert Armstrong via
cctalk <cctalk(a)classiccmp.org> wrote:
I have three DEC RA8x drives that have failed (all of them fault with
"spin error") because of bad photo-interrupter tachometer sensors. After
talking to a few friends, it sounds like this is a pretty common fault.
Photo sensors like this are fairly common, even today, but the specific
parts DEC used are weird and unobtainable.
Does the tachometer have to be accurate, or does it just have to indicate "spinning
fast enough" to satisfy the spin-up logic? In other words, does the actual laying
out of bits on the media depend on the tachometer?
If not, you could just use a 555 to generate a pulse train at the nominal frequency,
perhaps with a second 555 to delay that signal after power up to correspond to the typical
time it takes the drive to get up to speed.
paul
1:27 a.m.
Paul Koning <paulkoning(a)comcast.net> wrote:
Does the tachometer have to be accurate, or does it just have to indicate
"spinning fast enough" to satisfy the spin-up logic?
It's not just spin up - the firmware monitors the disk speed all the time
it's running.
But AFAIK the tachometer has nothing to do with how the bits are recorded
on the media. It's just there to be sure the disk is actually
rotating.
If not, you could just use a 555 to generate a pulse
train ...
Maybe, probably, but I'm not sure why you'd want to. That's way more
complicated than just fixing the tachometer sensor, and having the firmware
shut down the motor in the event the belt breaks or jams, or if the motor
brake (yes, it has a brake!) freezes, etc is an advantage. It's a fairly
big motor and a fairly (by PC standards at least) heavy disk - there's a lot
of mechanical energy there.
Bob
1:34 a.m.
On Nov 7, 2022, at 8:27 PM, Robert Armstrong
<bob(a)jfcl.com> wrote:
Indeed, but I was reacting to the statement that the DEC part is strange and hard to find.
So if that turns out to be barrier, faking the tach signal would be a way to make the
drive operational again.
Another approach would be a bit of mechanical work to fit a stock optical sensor. That
would depend on having access to the needed machining skills.
paul
Paul Koning <paulkoning(a)comcast.net>
wrote:
Does the tachometer have to be accurate, or does it just have to indicate
"spinning fast enough" to satisfy the spin-up logic?
It's not just spin up - the firmware monitors the disk speed all the time
it's running.
But AFAIK the tachometer has nothing to do with how the bits are recorded
on the media. It's just there to be sure the disk is actually
rotating.
If not, you could just use a 555 to generate a
pulse train ...
Maybe, probably, but I'm not sure why you'd want to. That's way more
complicated than just fixing the tachometer sensor, and having the firmware
shut down the motor in the event the belt breaks or jams, or if the motor
brake (yes, it has a brake!) freezes, etc is an advantage. It's a fairly
big motor and a fairly (by PC standards at least) heavy disk - there's a lot
of mechanical energy there. 
6:58 a.m.
-----Original Message-----
From: Paul Koning via cctalk <cctalk(a)classiccmp.org>
Sent: 08 November 2022 01:34
To: Bob Armstrong <bob(a)jfcl.com>
Cc: cctalk(a)classiccmp.org; Paul Koning <paulkoning(a)comcast.net>
Subject: [cctalk] Re: Replacement tachometer sensor for DEC RA80/1/2
drives ...
On Nov 7, 2022, at 8:27 PM, Robert Armstrong
<bob(a)jfcl.com> wrote:
Indeed, but I was reacting to the statement that the DEC part is strange Paul Koning <paulkoning(a)comcast.net>
wrote:
Does the tachometer have to be accurate, or does it just have to
indicate "spinning fast enough" to satisfy the spin-up logic?
It's not just spin up - the firmware monitors the disk speed all the
time it's running.
But AFAIK the tachometer has nothing to do with how the bits are
recorded on the media. It's just there to be sure the disk is
actually rotating.
If not, you could just use a 555 to generate a
pulse train ...
Maybe, probably, but I'm not sure why you'd want to. That's way more
complicated than just fixing the tachometer sensor, and having the
firmware shut down the motor in the event the belt breaks or jams, or
if the motor brake (yes, it has a brake!) freezes, etc is an
advantage. It's a fairly big motor and a fairly (by PC standards at
least) heavy disk - there's a lot of mechanical energy there. to find. So if that turns out to be barrier, faking
the tach signal would
be a way
to make the drive operational again.
The replacement board does not use the DEC part.
Another approach would be a bit of mechanical work to fit a stock optical
sensor. That would depend on having access to the needed machining
skills.
That is what this mod does
paul
dave
744
days inactive
745
days old
5 comments
4 participants
participants (4)
-
dave.g4ugm＠gmail.com -
Mike Katz -
Paul Koning -
Robert Armstrong