13 Jul
2007
13 Jul
'07
11:53 p.m.
From: Brent Hilpert <hilpert at cs.ubc.ca>
dwight elvey wrote:
Hi Bret
I recall that is was an HP counter. I don't recall the fastest count but
it was on the order of either 10KHz or 100KHz. I don't think it was
less then that but it may have been 1MHz. I recall that like many of the
HP counters of the time, the other decades were the inline 10 ea bulbs. I
don't recall if these were neon or incandescent but I think they were
neons driven by dual triodes. I do believe that the ring counter did
have a purplish glow.
Dwight
_________________________________________________________________
http://imagine-windowslive.com/hotmail/?locale=en-us&ocid=TXT_TAGHM_mig…
As a note on the dekatron, these were used in some of the early
HP counters. They were used as the first, or highest frequency,
state of the counters. They were quite fast!
Dwight
I beg to differ on this. I have a couple of dekatron counters and they
only seem to be good to some 10s of KHz (unless of course, you mean that
that
was pretty fast for the time). Interestingly, one of these counters uses
two
types of dekatrons, the only obvious difference between them being the
colour
of the glow and the part number. The ones in the slower speed decades are
the
usual neon glow, in the higher speed decade a purplish glow (xenon?). I
presume
the purplish glow gas ionizes faster than neon.
I should try to ascertain more accurately the max rate of the purplish deks
but haven't yet as one of the two in the unit is failing.
I haven't seen dekatrons in HP counters, I believe their first counter was
the
524A and, unless it was radically different than the 524B/C/D, didn't use
dekatrons. Certainly the rest of the 520 series was vacuum tube flip-flops
(AC-4 modules). 
14 Jul
14 Jul
2:06 a.m.
New subject: neon glow lamps (was TTL homebrew CPUs) in HP counters
dwight elvey wrote:
Hi Bret
I recall that is was an HP counter. I don't recall the fastest count but
it was on the order of either 10KHz or 100KHz. I don't think it was
less then that but it may have been 1MHz. I recall that like many of the
HP counters of the time, the other decades were the inline 10 ea bulbs. I
don't recall if these were neon or incandescent but I think they were
neons driven by dual triodes. I do believe that the ring counter did
have a purplish glow.
Dwight
I have the manuals for most of the early HP counters. I looked through
them and could not find
dekatrons used. The input goes directly to the counter units. The 523C
goes to 1.2 MHz.
In the 524B counter, there are two divide by 10 circuits built with tube
flip-flops. This
counter counts up to 10 MHz.
In the timebase circuit, they used a two tube circuit called a
phantastron to divide the 100KHz
crystal down. The circuits are decade dividers. In the 523C counters,
the tubes are
1/2 of a 6AL5 diode and a 6AS6. They work up to at least 100KHz.
A circuit and description can be found here:
http://www.dvq.com/images/phantastron.jpg
The odd thing about this is I swear I remember the ring counters also. I
don't remember a purple
glow, just the typical red-orange. I'm still looking through manuals to
see if I can find something.
Bob
2:31 a.m.
New subject: neon glow lamps (was TTL homebrew CPUs) in HP counters
On 13 Jul 2007 at 23:06, Bob Rosenbloom wrote:
In the timebase circuit, they used a two tube circuit
called a
phantastron to divide the 100KHz crystal down.
Which, of course, brings this to mind:
http://www.geocities.com/pdq340/humor/umac606.htm
Cheers,
Chuck
2:03 p.m.
New subject: neon glow lamps (was TTL homebrew CPUs) in HP counters
Chuck Guzis wrote:
Which, of course, brings this to mind:
http://www.geocities.com/pdq340/humor/umac606.htm
Utter genius. I need to print that one out for the Colossus guys... :-)
2:16 p.m.
New subject: vac tube clock rates / was Re: neon glow lamps (was TTL homebrew CPUs) in HP counters
Bob Rosenbloom wrote:
dwight elvey wrote:
The remaining 6 of the decades in the 524's are built with vac-tube flip-flops
as well. Pardon the detail following but there is an interesting point (IMHO):
by count rate there are 3 types of counters in the 524's, the lower speed 6
decades (<=100KHz) are a typical 4 flip-flop circuit built from 4 duo-triodes,
the 1 Mhz counter is similar but with some clamping diodes and lower
resistances, however the 10Mhz stage is about 3 times as complex: 13 tubes
(not duo-triodes), a bazillion clamp diodes, a very different circuit
arrangement and even count sequence.
(FWIW: http://www.cs.ubc.ca/~hilpert/e/edte/HP520.html for my own take on the series.)
It's a good example of the limits of vacuum-tube technology which applied to
the computers of the day (at least with typical tubes in the 50s). Going
beyond a few MHz became infeasible. For references I have at hand the fastest
clock rate I see is 2.5 MHz for BINAC, a lot of the tubes machines seem to
have run at around a MHz or less. Anybody know what the basic clock rates of
the IBM 709 or Sage processors was? (some of the last big tube machines).
Hi Bret
I recall that is was an HP counter. I don't recall the fastest count but
it was on the order of either 10KHz or 100KHz. I don't think it was
less then that but it may have been 1MHz. I recall that like many of the
HP counters of the time, the other decades were the inline 10 ea bulbs. I
don't recall if these were neon or incandescent but I think they were
neons driven by dual triodes. I do believe that the ring counter did
have a purplish glow.
Dwight
I have the manuals for most of the early HP counters. I looked through
them and could not find
dekatrons used. The input goes directly to the counter units. The 523C
goes to 1.2 MHz.
In the 524B counter, there are two divide by 10 circuits built with tube
flip-flops. This
counter counts up to 10 MHz. 
2:35 p.m.
New subject: vac tube clock rates / was Re: neon glow lamps (was TTL homebrew CPUs) in HP counters
It's a good example of the limits of vacuum-tube
technology which applied to
the computers of the day (at least with typical tubes in the 50s). Going
beyond a few MHz became infeasible.
Without going to the old industry rags (I am not home now anyway), I
would bet that the upper limit on computer tube operation is due to
the interelectrode capacitance and spacing issues (the actual distance
between the cathode and plate - it become criticals when you see that
the electron transit speed is actually not that fast). I would also
bet that the speed could have been increased if the tubes shrank -
using subminis instead of minis. Speed would probably go up, but cost
would also skyrocket.I doubt any commercial maker would have wanted to
pay that price.
--
Will
2:47 p.m.
New subject: vac tube clock rates / was Re: neon glow lamps (was TTL homebrew CPUs) in HP counters
On 14 Jul 2007 at 14:35, William Donzelli wrote:
Without going to the old industry rags (I am not home
now anyway), I
would bet that the upper limit on computer tube operation is due to
the interelectrode capacitance and spacing issues (the actual distance
between the cathode and plate - it become criticals when you see that
the electron transit speed is actually not that fast). I would also
bet that the speed could have been increased if the tubes shrank -
using subminis instead of minis. Speed would probably go up, but cost
would also skyrocket.I doubt any commercial maker would have wanted to
pay that price.
When I used to shopping on Chicago's Michigan Avenue "Surplus Row", I
used to see lots (boxes and boxes) of PCBs with submini wire-lead
dual triodes on them (usually about 6 per card). I bought them and
used the tubes for HF operation (very handy in a cascode
configuration) and lots cheaper than Nuvistors. Since most of these
stores dealt in defense surplus, I suspect these were part of some
military computer as most of the boards were wired as flip-flops.
Cheers,
Chuck
15 Jul
15 Jul
2:06 a.m.
New subject: vac tube clock rates / was Re: neon glow lamps (was TTLhomebrew CPUs) in HP counters
From: "Chuck Guzis" <cclist at
sydex.com>
On 14 Jul 2007 at 14:35, William Donzelli wrote:
These might have been use for encryption. They used a lot of
LFSRs in these. That would be mostly flops and little else.
Dwight
_________________________________________________________________
Need a brain boost? Recharge with a stimulating game. Play now!?
http://club.live.com/home.aspx?icid=club_hotmailtextlink1
Without going to the old industry rags (I am not
home now anyway), I
would bet that the upper limit on computer tube operation is due to
the interelectrode capacitance and spacing issues (the actual distance
between the cathode and plate - it become criticals when you see that
the electron transit speed is actually not that fast). I would also
bet that the speed could have been increased if the tubes shrank -
using subminis instead of minis. Speed would probably go up, but cost
would also skyrocket.I doubt any commercial maker would have wanted to
pay that price.
When I used to shopping on Chicago's Michigan Avenue "Surplus Row", I
used to see lots (boxes and boxes) of PCBs with submini wire-lead
dual triodes on them (usually about 6 per card). I bought them and
used the tubes for HF operation (very handy in a cascode
configuration) and lots cheaper than Nuvistors. Since most of these
stores dealt in defense surplus, I suspect these were part of some
military computer as most of the boards were wired as flip-flops.
3:06 a.m.
New subject: vac tube clock rates / was Re: neon glow lamps (was TTLhomebrew CPUs) in HP counters
On 14 Jul 2007 at 23:06, dwight elvey wrote:
These might have been use for encryption. They used a
lot of
LFSRs in these. That would be mostly flops and little else.
Dwight
The triodes were attached with clips to anodized aluminum "wings" on
either side of the PCB, with small front panel with a couple of test
points on it. They were also a handy source of silvered mica caps.
Maybe 8-10 inches wide by 3/4 inch tall by 9 inches deep. No
manufacturer's name that I recall.
After over 40 years, that's all I can remember about them. Back
then, I think I paid about a buck apiece for each board. 6.3vac
heaters.
Cheers,
Chuck
14 Jul
14 Jul
2:43 p.m.
New subject: vac tube clock rates / was Re: neon glow lamps (was TTL homebrew CPUs) in HP counters
On 14 Jul 2007 at 11:16, Brent Hilpert wrote:
Anybody know what the
basic clock rates of the IBM 709 or Sage processors was? (some of the last
big tube machines).
0.667 MHz.
--Chuck
5:39 p.m.
New subject: vac tube clock rates / was Re: neon glow lamps (was TTL homebrew CPUs) in HP counters
On Sat, 14 Jul 2007, Chuck Guzis wrote:
On 14 Jul 2007 at 11:16, Brent Hilpert wrote:
It's difficult to find any info about clock speed for SAGE, but the
core memory access time was 6 microseconds:
http://ed-thelen.org/comp-hist/BRL61-a.html#SAGE
Mike Loewen mloewen at cpumagic.scol.pa.us
Old Technology http://sturgeon.css.psu.edu/~mloewen/Oldtech/
Anybody know what the
basic clock rates of the IBM 709 or Sage processors was? (some of the last
big tube machines).
0.667 MHz.
6221
days inactive
6222
days old
10 comments
7 participants
participants (7)
-
bobalan＠sbcglobal.net -
cclist＠sydex.com -
dkelvey＠hotmail.com -
hilpert＠cs.ubc.ca -
julesrichardsonuk＠yahoo.co.uk -
mloewen＠cpumagic.scol.pa.us -
wdonzelli＠gmail.com