EEVblog Electronics Community Forum
Electronics => Projects, Designs, and Technical Stuff => Topic started by: Jester on April 09, 2019, 03:53:36 pm
-
I have an old school Tektronix TM503 timer module that comes in handy once in a while for long duration timing (like an hour or a day). Uses a bunch of 74xxS logic and it runs quite hot. The original "S" parts are pretty power hungry.
What series would be the best to replace these with?
LS will only sink 8mA vs. 20mA for the "S", so this may or may not be a problem depending on the particular circuit (fanout)
ACT have good sink and source capability, improved logic levels and similar propagation delay, and they are also inexpensive.
-
I you replace *all* of the ICs from 74S to 74ACT, the respective input currents will also be reduced, and therefore the the 74ACT's lowered sink current should suffice.
Having said this.....Consult the datasheet that the reduced input current is proportional to the reduced output drive.
EDIT: I would assume that the original components are all thru hole. Will you obtain all the replacements in thru hole also?
-
I have done the same thing to a number of my TM500 series instruments. FACT or ACT has about twice the delay of schottky TTL. AS (advanced schottky) is a tiny bit faster at half the power and FAST (fast advanced schottky TTL) is a tiny bit slower at 1/4 the power.
-
Most of the time you can just replace S with LS, though the only times you get issues are things that do odd things, like having RC networks to do delays, or oscillators. If you do all at one there is no problem with fan outs or fan in, as the families all have the same fan out and fan in irrespective of family, provided you stay in the family. ACT will also work, just you might need to buy a good number of DIL to SMT adaptor boards, a few dozen leaded components to use as lead donors, and a few hours of soldering to get them all done.
-
I have done the same thing to a number of my TM500 series instruments. FACT or ACT has about twice the delay of schottky TTL. AS (advanced schottky) is a tiny bit faster at half the power and FAST(fast advanced schottky TTL) is a tiny bit slower at 1/4 the power.
David,
Do you recall what series you used?
Mostly socketed DIP packages in my case. I might just swap out what ever is available in DIP and see what happens. Hopefully the temperature drops a bit, the power supply relaxes and everything keeps working for a few more decades;-)
-
Most of the time you can just replace S with LS,
Tektronix never used S where LS or standard TTL would do. LS can usually replace standard TTL but not always. I think they dropped H (high power TTL) in favor of S as quickly as possible simply because of availability.
L (low power TTL) may have been used in the crystal oscillator circuit. If so, do not change it.
Do you recall what series you used?
I did not replace many schottky TTL chips but my preference was to use FAST. Neither AS nor FAST are very common so I used whatever was available.
Mostly socketed DIP packages in my case. I might just swap out what ever is available in DIP and see what happens. Hopefully the temperature drops a bit, the power supply relaxes and everything keeps working for a few more decades;-)
Most of the benefit will come from replacing standard TTL with LS (low power schottky) TTL because there are so many more of them. LS is actually faster but only has half the drive of standard TTL so this is not always feasible.
Standard TTL can be replaced with FAST (fast advanced schottky TTL) with no loss in drive and some power conservation. ACT can also replace standard TTL while maintaining high output drive; AHCT and HCT cannot.
I really like FAST as a high power and high speed alternative to standard TTL and LS but not every function was produced.
-
A couple of points to bear in mind when replacing one logic family with another.
Firstly a lower output impedance might just create problems with ringing or current spikes.
Secondly, faster isn't always better. In some designs a flip-flop's hold time (cf setup time) is on the critical path, and a faster driver would erode that margin. Whether that is significant depends on the detailed circuit, of course.
Having noted that, David Hess' approach is sane :)
-
I was looking at some of the parts used and they're a blast from the past. The Motorola MC4035 appears to be a single sourced 4 bit OC latch with an Icc of 40-50 mA. The Fairchild 9390/7490 decade counters have an Icc of 30-40 mA. Digikey has 7490 (!!) and 74LS90 listed. I don't see any pin compatible, lower power devices with a quick search. I think the reference clock chain (now 7490s) might be replaced by using some 74HC390 dual decade counters on some artfully arranged plug in headers and jumpers, but some detailed diving into the circuit would be needed first. Also note, the first two parts don't have their Vcc and Gnd pins in the corner.
I built a 4 digit frequency counter in the 70's with TTL MSI. It was simpler than the DC 503 but it still used at least an amp at 5 V. I also built a DMM that could operate from batteries. To save power, I found on both I could reduce Vcc to the 9368 seven segment display drivers with a series diode or two. Since they worked at less than kHz rates when driving the display they didn't need the full 5 V supply.
Cheers,
-
The Fairchild 9390/7490 decade counters have an Icc of 30-40 mA.
That one is actually in my notes; you must be going through the same process I did. Unfortunately the 7490 was only made available in TTL, L, and LS versions. Be careful; I think the L version has an alternative pinout.
The Signetics 8290 decade counter is TTL 40 MHz and 48.5 mA. The Signetics 8292 decade counter is the same thing but LS TTL 5 MHz and 13.1 mA. They can be replaced by the 74196 (faster and TTL, LS, and S versions) and 74176 (slower and TTL only).