Taylor Edge Nixie Clock Kit

[dave....@comcast.net]

I'm hoping that someone out there can help me with a problem that I am having with a Taylor Edge clock kit that I built. After completing the kit and plugging it in it seemed to be working fine. Then I noticed that the 10s second display tube and the 10s minute display tube were not counting correctly. They would count from 2 to 3 to 4 to 5 but then it would go back to 3 then 4 then back to 2 and then start the sequence over again. The 1s second and 10s minute tubes would count fine. I asked a friend of mine that knows a lot more about this stuff than I do and he recommended that I try disconnecting resistors R13 and R15 (both 240 ohm) from the time setting circuit to see what would happen. This fixed the counting problem but now I cannot set the time. I can't believe that I am the only person that has had this problem and am hoping someone out there knows the answer. I am attaching the schematic.

Thanks

Dave

[gregebert]

Clocking problem. I've been looking at the peculiar buffering on the clock signals (7417 open-collector buffer), and cant come up with a good reason why R11, R13, and R15 are present. I'd be especially concerned about unstable operation, as this circuit has positive feedback. 

Are you using the exact parts,which seem to be 7400-series TTL, or did you  use something similar, such as 74LS, 74LV, 74HC ? I suspect the circuit is very fussy about threshold values and output-drive-strength, so swapping devices that are functionally, but not parametrically equivalent, could cause problems.

If you have a digital scope (you will need to capture the signal), best to start looking at clock-edges for odd behavior (ringing/glitches, incorrect amplitudes). 

[Nick]

I suspect they're there to "latch-up" the buffer to debounce the switches - pulling the input to ground will "de-latch" it, i.e. it's an ideal toggle switch or latching bistable.

Nick

[dave....@comcast.net]

Your question about the TTL type was discussed between my friend and I. He asked me why I used 74LS instead of 7400. I told him that 74LS00, 74LS90, and 74LS92 is what was on the BOM for the kit. He also thought that a 7400 may work better. I will try that first and see what happens. Will let you know what happens. Thanks for the advise.  

[Nick]

If the BOM specified LS, use LS. The Taylor stuff is good - it's more likely that you have an assembly error.

Randomly changing stuff to see what happens is not a good approach.

I'd start by very thoroughly checking that you have the right components in the right places, that you don't have any dry joints or solder bridges. Put the resistors you took out back in - they are there for a reason as I mentioned previously.

My money is on you having made a mistake somewhere - problem solving requires a methodical approach, following the signal chain from source, in this case from the 60Hz take off from the power input to the board...

Do you have a 'scope?

Nick

[Nick]

...Also, reading your problem description isolates the issue to the 10s digits.

The 10s are driven by 7490s whereas the units are driven by 7492s. Do you have the right chips in the right places?

I'd also try swapping round the 74141s between the 10s and units to see if the problem moves with them - 74141s can fail....

Isolate the problem.

Nick

[Tidak Ada]

It is also possible one of the 7490 or 4717 chips isn't functioning correct or damaged or a bad soldering. Check your soldering work with a magnifier glass

eric

-----Oorspronkelijk bericht-----
Van: neoni...@googlegroups.com [mailto:neoni...@googlegroups.com] Namens Nick
Verzonden: donderdag 9 maart 2017 4:07
Aan: neonixie-l
Onderwerp: [neonixie-l] Re: Taylor Edge Nixie Clock Kit

--
You received this message because you are subscribed to the Google Groups "neonixie-l" group.
To unsubscribe from this group and stop receiving emails from it, send an email to neonixie-l+...@googlegroups.com.
To post to this group, send an email to neoni...@googlegroups.com.
To view this discussion on the web, visit https://groups.google.com/d/msgid/neonixie-l/c6ffdd07-5d4d-4de6-8afc-37d5e7f1da8e%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[JohnK]

Very very good advice from Nick in these two posts.

Changing like-for-like or interchanging if easy is valid - modifications
aren't. [Unless you deduce something from a *temporary* mod to allow you to
fix the original problem.]
I have had to deal with the fallout from bad approaches: "We" discovered
that, in the Chinese factory, there was an 'engineer' modifying the units
that failed test until they worked or passed. Things like changing the value
of a resistor in a voltage divider that supplies a ref voltage to a
comparator making up for an IC with a leaky input ( or a solder bridge to a
different part of the circuit as in one of the customer returns that I
analysed). Eeek!!

John K

[David Speck MD]

Along the same line, is there a chance that you have a counterfeit 7490
chip? A lot of chips coming from Asia are marginal manufacturing
rejects or remarked circuit pulls that may have been mid-labeled.

Dave

[Terry S]

First we need some clarification of the problem, as your original post contains a discrepancy. At one point you say the 10 minute tube is not counting correctly, then later you say it is fine. This is probably a typo, but let's hear the answer.

Nick, the 10s are driven by 92's and the 1s by 90's. Another minor correction for clarity.

Here is where I would break out the logic analyzer and hang some probes, watch the count sequence. This appears to be a case of the counters not resetting properly as they roll over from 5 to 0.  This could be for a number of reasons, maybe the parts are faster than the used to be, subtle changes in setup & hold times can occur with die changes, shrinks, technology changes, etc. Maybe a ringing clock that never used to be a problem is now..

USB based logic analyzers are handy tools, and have gotten pretty cheap.

Check the counting with & without the 74141s installed. If they are loading the counter outputs, they could affect the count sequence, as some outputs are also used internally on the counters to sequence the F/Fs.

Terry

[dave....@comcast.net]

Just wanted to update everyone on where we are at on this project. One of the first things that my friend Mike questioned was why are we using 74LS series instead of 7400. I have order new ICs for everything except the 141s. Here is the last update that I received from him.

  

Ok, did some quick calculations, and the drive current needed to drive the 7417 from the output of the counter with the 360 ohm resistor to VCC, you need 14ma.  The current output of the 74LSxx counter is only 8ma, and the 74xx is 16ma.  When and if the 7417 see's a good logic low on it's input, it's output will switch low pulling the 510 ohm low, 9ma needed, helping drive the counter output low.  The 74LS17 can drive 30ma, so that can help once it can see a good logic low (.8V) on the input.  So, if we increase the resistance of the 360 ohm resistor to 625 ohms, that may be a place to start.  These are rough calculations, but I would go back to using the 74xx's instead of the 74LSxx's this would have been better.

Mike

[Dekatron42]

What happens if you set all three time switches and the halt/set switch in the SET position and then push the INCR button, do all counters advance one step each so that they count properly or do soem still count in the strange sequence?

I would also make sure that VCC and GND are present on all places where they should be present, like on all IC pins and on all component pins - having a loose VCC or GND pin can wreak havoc on a system and is very hard to find by reasoning or by other tests.

/Martin

[JohnK]

OK, so now I looked at the schematic you posted [it is a slightly changed TI ap it seems].

I took it at face value that the BOM was correct, but I don't see one in that pdf you attached.

 

The schematic uses 7400 series.

 

There are  "Changes to original..." at top left of sheet 4.

In those changes I see no mention of LS being considered.

 

I agree that it IS a good approach to check the design calculations - especially in this case where the documentation apparently mismatches.

 

A question then... are the resistors being discussed  ( sh1  R1, R2, R3 ) actually  the talked about 360 ohms? Have their values been changed on the board or in the BOM?

 

 

John K.

----- Original Message -----

From: dave....@comcast.net

To: neonixie-l

Sent: Friday, March 10, 2017 10:41 PM

Subject: [neonixie-l] Re: Taylor Edge Nixie Clock Kit

--

You received this message because you are subscribed to the Google Groups "neonixie-l" group.
To unsubscribe from this group and stop receiving emails from it, send an email to neonixie-l+...@googlegroups.com.

To post to this group, send email to neoni...@googlegroups.com.
To view this discussion on the web, visit https://groups.google.com/d/msgid/neonixie-l/8f1bbc47-9661-4e52-9a19-54b21668a3a3%40googlegroups.com.

[gregebert]

Take a close look at the signal amplitudes (best to use a scope); if the resistor values are causing the logic gates to be overloaded [either due to wrong IC, wrong logic family, wrong resistor value, solder short/open] , you will see incorrect levels. Generally, logic-0 will be 200mV or less (despite the specs being Vol(max)=0.4), and logic-1 around 4V (despite spec being Voh(min)=2.4). Remember - clock signals are critical; they must have clean edges and should swing rail-to-rail. Anything other than this should be suspect.

In my previous life as a technician, one of the first things we did when debugging a large PCB was to hunt-around ICs for bad signal amplitudes. It took less than a minute to probe a large section, and that method found many problems (wrong IC, backwards IC, bent pin, solder-bridge, bad PCB trace, bad socket). In *rare* cases, the actual IC was bad, which seems counter-intuitive until you realize these devices are thoroughly tested by the manufacturer.

[dave....@comcast.net]

Here is a copy of the BOM

On Tuesday, March 7, 2017 at 6:37:54 PM UTC-6, dave....@comcast.net wrote:

[JohnK]

Well, I was waiting off for other comments to surface or for the kit supplier to pipe up.

 

 I see a few choices:-

- contact the supplier for comment

- measure waveforms/ logic levels and act accordingly

- try specially selecting 74LSxx that work in the location [might be temperamental with age, temperature etc]

- try using the components shown on the schematic, ie 74xx series. BUT, check power supply ratings etc first

 

If it were mine I would want to measure the goings-on with an oscilloscope.

 

John K.

 

 

 

 

 

 

----- Original Message -----

From: dave....@comcast.net

To: neonixie-l

Sent: Saturday, March 11, 2017 5:28 AM

Subject: [neonixie-l] Re: Taylor Edge Nixie Clock Kit

--

You received this message because you are subscribed to the Google Groups "neonixie-l" group.
To unsubscribe from this group and stop receiving emails from it, send an email to neonixie-l+...@googlegroups.com.
To post to this group, send email to neoni...@googlegroups.com.

To view this discussion on the web, visit https://groups.google.com/d/msgid/neonixie-l/6f0b3500-cd4f-4029-94bb-5f3ace5102d2%40googlegroups.com.

[dave....@comcast.net]

Sorry that it took me so long to get back about this clock problem. We have the clock work perfectly now. I have to give my friend Mike all the credit for figuring this out. It appears that the original clock design schematic calls for 7400 series ICs but the BOM calls for 74LSXX. Mike has a IC spec. book and after comparing the differences in the ICs we found that the 74LSXX have an output of 8ma and the 7400 series have an output of 16ma. We believe that the original design was for 7400 but in shopping for them found that there are very few suppliers anymore. We think that this is why they changed the BOM to 74LSXX. With the smaller output from the 74LSXX IC and the amount of resistance in that circuit the current was marginal and not enough current to drive the other ICs correctly in the circuit. We ended up removing R1 and R3 (360 ohm) and leaving everything else the way it was. It works perfect now and all of the setting switches work fine also. Thanks for all the feedback from everyone

Dave    

On Tuesday, March 7, 2017 at 6:37:54 PM UTC-6, dave....@comcast.net wrote:

[JohnK]

Good outcome. I wondered what had happened.

Saw what you meant as soon as I looked at the docs. At least you know what tweaks to do if anymore funnies turn up [ageing, temperature etc].

 

You can often scavenge the old parts from recycling centres if they haven't actually chewed up the boards for metal extraction.

I am sure that all the boards we see leaving here for the Philipines get components pulled and sold on ebay.

 

I regret putting off getting a bigger selection of spares for all my 'interesting' old junk - like Data General Novas.

 

John K.

----- Original Message -----

From: dave....@comcast.net

To: neonixie-l

Sent: Friday, March 31, 2017 11:55 AM

Subject: [neonixie-l] Re: Taylor Edge Nixie Clock Kit

--

You received this message because you are subscribed to the Google Groups "neonixie-l" group.
To unsubscribe from this group and stop receiving emails from it, send an email to neonixie-l+...@googlegroups.com.
To post to this group, send email to neoni...@googlegroups.com.

To view this discussion on the web, visit https://groups.google.com/d/msgid/neonixie-l/51ff2297-8b93-4126-a14c-6a31687c11a1%40googlegroups.com.

[threeneurons]

I'm a little late to the party, and the discussion between old 74xx versus 74LSxx caught my attention. In most cases, I'd favor the newer 74LS over the original 74xx. But the resistor values (360 ohm) paint you into using the older series, because 5V / 360ohms gives you 13.9mA. The value seems to have been chosen because of the set switch resistor values.Of course as the OP mentioned, removing all the 360 ohm resistors lets the unit operate properly. The set switches may work adequately, too.

Personally, I prefer 4000 series CMOS, for this kind of thing.

Good to see that the OP, and his buddy, figured it out.

[chuck richards]

Yep, those old plain 7400 series TTL ics are little juice-hogs!

Changing that whole thing over to 74LS is a good idea.
The plain old 7400 series TTL works ok, but it hogs a lot
more current, and it's also much more susceptible to noise
and transient spikes. 74LS series is a vast improvement!

But, you are correct in your observation that when changing
from plain 74XX to 74LSXX, that particular attention must be
paid to current availability.

Swapping out those resistors sounds reasonable.

>>To post to this group, send an email to neoni...@googlegroups.com.

>>To view this discussion on the web, visit
>https://groups.google.com/d/msgid/neonixie-l/51ff2297-8b93-4126-a14c-
>6a31687c11a1%40googlegroups.com.
>>For more options, visit https://groups.google.com/d/optout.
>>

$4.95/mo. National Dialup, Anti-Spam, Anti-Virus, 5mb personal web space. 5x faster dialup for only $9.95/mo. No contracts, No fees, No Kidding! See http://www.All2Easy.net for more details!

[chuck richards]

360 ohms is way too little resistance for *any* sort of
TTL pullup resistor! That apparently got confused with
the standard *pull down* resistor used to keep an unattended
TTL input *low*. 360 ohms for a TTL pulldown is just right.

2.2k ohms is the standard value to use as a TTL pullup.
That goes for regular old 7400 series TTL, as well as 74LS
series.

The old military practice was to always use 1k ohm as the
standard TTL pullup resistor value. Note that even when using
that 1k pullup resistance, that all we are asking of any TTL
gate which is pulled up by this, is to be able to provide 5
milliamperes
in order for for the output to go low.

Even that value of 1k ohm would work ok with either plain 7400 or
74LS.

2.2k is much more like it, however.
At that value, all we are asking is 2.3 mA, and that works fine.

360 ohms is just simply too close to being a piece of wire!!

(This is all explained in great detail on page 12 of Don Lancaster's
famous TTL Cookbook)

---- Original Message ----
From: neoni...@googlegroups.com
To: neoni...@googlegroups.com
Subject: [neonixie-l] Re: Taylor Edge Nixie Clock Kit
Date: Thu, 30 Mar 2017 19:38:26 -0700 (PDT)

>
>
>
>I'm a little late to the party, and the discussion between old 74xx
>versus
>74LSxx caught my attention. In most cases, I'd favor the newer 74LS
>over
>the original 74xx. But the resistor values (360 ohm) paint you into
>using
>the older series, because 5V / 360ohms gives you 13.9mA. The value
>seems to
>have been chosen because of the set switch resistor values.Of course
>as the
>OP mentioned, removing all the 360 ohm resistors lets the unit
>operate
>properly. The set switches may work adequately, too.
>

><https://lh3.googleusercontent.com/-k_4j6X4xaEs/WN3AZVxQbgI/AAAAAAAAY
>lU/Wcu3xmkZidsfPKYgHCTU5VlWEi_nSXiWACLcB/s1600/Taylor_Clock.jpg>

>
>Personally, I prefer 4000 series CMOS, for this kind of thing.
>
>
>Good to see that the OP, and his buddy, figured it out.
>

>--
>You received this message because you are subscribed to the Google
>Groups "neonixie-l" group.
>To unsubscribe from this group and stop receiving emails from it,
>send an email to neonixie-l+...@googlegroups.com.
>To post to this group, send an email to neoni...@googlegroups.com.
>To view this discussion on the web, visit

>https://groups.google.com/d/msgid/neonixie-l/6d9e738a-b183-4284-846e-
>d1e2091fefa1%40googlegroups.com.

[JohnK]

Yeah Chuck.. but the interesting thing is that it is supposedly very closely
based on a TI ap note or some such.
I was looking forward to some input from the seller. But maybe this is
public and wasn't "addressed" to him.
Still think it is/was in his interest to comment though.

John K

> https://groups.google.com/d/msgid/neonixie-l/380-220173531114249955%40all2easy.net.

[threeneurons]

On Friday, March 31, 2017 at 4:42:53 AM UTC-7, Chuck wrote:

360 ohms is way too little resistance for *any* sort of
TTL pullup resistor!  That apparently got confused with
the standard *pull down* resistor used to keep an unattended
TTL input *low*.   360 ohms for a TTL pulldown is just right.

2.2k ohms is the standard value to use as a TTL pullup.
That goes for regular old 7400 series TTL, as well as 74LS
series.  

The old military practice was to always use 1k ohm as the
standard TTL pullup resistor value.  Note that even when using
that 1k pullup resistance, that all we are asking of any TTL
gate which is pulled up by this, is to be able to provide 5
milliamperes
in order for for the output to go low.

Even that value of 1k ohm would work ok with either plain 7400 or
74LS.

2.2k is much more like it, however.
At that value, all we are asking is 2.3 mA, and that works fine.

360 ohms is just simply too close to being a piece of wire!!

(This is all explained in great detail on page 12 of Don Lancaster's
famous TTL Cookbook)

I think he's stuck with the low value resistors, because he's making Schmitt triggered gates out of 7417s. The feed back resistor has to work both as a pulldown, and pullup. As noted, TTL source and sink, are very asymmetrical. 

This is one, of several reasons, I prefer, 4000 series CMOS over 74 TTL, when it comes to older logic families. CMOS is symmetrical. It has high impedance inputs, and you're not stuck with a 5V supply. On my "new" thermometer kit, I run the CMOS on 12V. Here is the set switch scheme on my old nixie clock board:

The "seconds-to-minutes" and "minutes-to-hours" pulses are passed thru 1000pf caps, feeding 4.7K loads, that normally sit at 5V. It has to use a 5V supply due to the 74141 drivers, also used. Short pulses are seen at the respective inputs, for minute, and hour, "bumps". Setting is done thru 4093 Schmitt gates, which are made "open drain", by using 1N914 diodes, in their output legs. Since its 4000 CMOS, a logic-0 is anything under 1/3 supply, or ~1.5V or less. A logic-1 is anything over 2/3 supply (>3.5V). The 0.7V diode drop is a non-issue, here. Punching the "hours" or "minute" button, increases the count by one, per depression. No goofy accelerated clock scheme. This would be much more difficult to implement in TTL.