TSP #337 - How Can HP 5245L Nixie Tube Frequency Counter Measure DC Voltage? Theory, Teardown & Demo

追加: 2026-05-23

[00:00:07]Hi, welcome to Signal Path. In this episode, we're going to return to this really beautiful old instrument and do something new with it. This is, of course, the HP 5245L electronic counter, a really famous counter, specifically recognizable by its beautiful Nixie display set. Now, several years ago, I did a full tear down on this, did some repairs on it, talked about how it works, and we even did something fairly unusual with it where I used a bunch of external synthesizers to turn it into a clock. Essentially generating frequencies that would represent the clock symbols directly here on the display, which was quite fun just from an experiment. It was probably the most complicated Nixie tube clock anyone has ever put together. But today, I want to do something different. I want to turn it into a voltmeter. Now, this module up here, in fact, does that. These modules can be replaced. You can see this is for example a frequency extender module.

[00:00:52]There's a preset module which is a frequency divider. This might be brand new and it could be one of the rarest modules for this instrument. And I was lucky to eventually find this after many years of it having it on my eBay watch list. But how does this actually work?

[00:01:05]How does this turn this frequency counter into a multimeter? Well, let's think about this and maybe brainstorm and figure out how this could potentially be doing that. So now let's be creative. Let's think of a few different ways we can use a frequency counter to measure a DC voltage. We know how a frequency counter like this works.

[00:01:22]If you apply a single tone at its input, it's going to compare that with an internal stable reference, is going to tell you what the frequency of the input is. So why don't we use it exactly like that? Let's take a voltage controlled oscillator, take the oscillator output directly connected to the input of our frequency counter and then use the DC input voltage, the voltage that we want to measure and connect it to the tune voltage of the voltage controlled oscillator. So what would happen? Well, depending on the voltage that you put in, the voltage control oscillator will go to a new frequency. That frequency would go into the frequency counter and then it will be displayed over here. And if we know the exact relationship between the DC input and the frequency output of our VCO, then we will be able to program that into the frequency counter and that will be a voltage representation of what is at the input.

[00:02:06]Well, fundamentally and theoretically this works perfectly fine. But this actually introduces a lot of different complexities. First of all, you need to have a VCO with a really wide tuning range. Not to mention, you would have to have a very very good linear tuning range where the DC input voltage and the output voltage frequency would have to be really predictable and as linear as possible. Remember that there is no calibration or there wasn't any fancy memory and things like this where you could plot and characterize this really well and then use a lookup table for example. And then of course these things are very sensitive to temperature and they're going to move around. They're sensitive to aging. Just building a stable voltage reference inside of an instrument like this is often oven controlled. We've talked about various oscillators quite a bit. So even though this works in theory is actually a really difficult way to do this and not a very good way. But funny enough, this is exactly what I did to turn this into a clock. In one of the previous videos, I used a programmable synthesizer which would then show the displays of the clock over here by programming it to a new value every second. And that was fine for a novelty, but that's not really a good way to do that. So we should scratch this. So let's think a little bit deeper about how this frequency counter works internally and some of its other capabilities. In reality, internally it has a counter, a totalizer which can have a start and stop condition. In fact, it uses that in combination with its own internal counter based on its own oscillator to figure out what the input frequency is.

[00:03:29]That's a really rough way of describing how this works, but it is suffice to say that if you have a counter now, you can do some other interesting things with it. Now remember when you start this count and stop this count, it has a very precise step because it's set by the internal oven control oscillator that's inside of this instrument. So now that we have a precision counter that we can start and stop anytime we want, why don't we wrap some circuit around that and take advantage of this counter itself. Let's think about a situation like this. This is a very simplified diagram, but it will get the point across. Imagine that your DC input comes over here and we apply that DC input to a comparator. And then on the other side of the comparator, we build a really simple integrated circuit. Now we have a start count and we connect the stop count to the output of the comparator.

[00:04:11]So imagine that at any given time, we can start this counter. So what would happen? Well, at the start of the counter, we're going to make sure this capacitance is fully discharged by connecting it through this switch to zero. So the capacitor starts at voltage zero. That would be somewhere over here.

[00:04:25]At that point, we get the counter to start counting. Then we close this circuit pushing a constant current I through this capacitor. Well, what happens when you push a constant current into a capacitor? You get an integrator whose voltage goes up exactly linearly.

[00:04:39]And it's only a function of I and C. So, you get this nice ramp starting to go up. So, at some point, this ramp is going to charge this capacitor to a voltage. And that voltage is going to reach the DC input. That would be this threshold over here. At that exact moment, this comparator would flip and it will tell this stop counter to stop counting. So then now you have this time between the start and stop which is counting very precisely and you know exactly the slope of this line that slope is proportional to the value of I and C. Now if you can build a really nice and stable current source which we can and if you have a really low leakage capacitor which we do then you can be able to create the slope precisely and it can be temperature stable. It can be really well controlled and it can have different slopes depending on how fast you want it to work. So then at that point the counter stops. So you know the frequency of the counter that's inside of the instrument. You know the value of I you know the value of the C. And then by figuring out how much count you had, you know exactly what your DC input voltage was. Well, we kind of reinvented an integrating ADC. In fact, there are many different versions of this. There's a dual slope version of it where you charge a capacitor and then discharge the capacitor with a known reference voltage. This is a super super simplified version of this, but it's really easy to implement. Importantly, it is based on components that were available at a time an instrument like this was built. Of course, there are many different kinds of ADCs, and I have covered quite a few of them on the channel before in a lot of details. So, this type of integrating ADC should be pretty easy to build. That's roughly what that module does. Now, the module is a bit more complicated, and I'll talk about some of its complications, but you can see that this is such a better way of using this instrument is with this built-in counter and a simple IC component to be able to measure voltage.

[00:06:18]We should be doing it this way. And of course, as to be expected, the actual instrument is quite a bit more complicated and with a lot of elegant circuitry, especially because of the limitations of the components they had available to them back then. Now this unit does not have autoscaling which means in order to select the scaling which the voltmeter works there is an input attenuator and a voltage switch that you have to manually select and there's also an 8vt reference built in which we need to use to calibrate the circuit. There's also a zero calibration which we will see when we play around with it. Now there is a RAM generator just like I described earlier but it is applied to more than one place. It is applied to an input comparator as well as a ground comparator because it also has to compare with the reference ground to know where those transitions actually happen. And you can see that ultimately the start counter and stop counter signals are generated from these two comparators as to be expected. There's also a polarity indicator in the front because you don't know if the voltage is going to be positive and negative and it figures it out depending on which order these pauses arrive. So there's some cleverness in there as well. So that's really the overall block diagram, but of course there's a lot more going on. And here's a much more detailed block diagram of the instrument. It's going to take forever to go through everything, but I do want to highlight a few important things here. So here's the input comparator section. You can see that the input signal once it arrives, it is going through an input comparator diode here that's going to then ultimately generate our stop signal. At the same time, the ramp generator here in the middle is pretty important because it applies the signal both at the top comparator and at the bottom comparator. And here's the ground comparator again comparing the ramp coming in now with a zero volt. And you can adjust that. It's going to give you that zero threshold which is pretty important. It is one of the front panel accessible potentiometers. There's also the 8volt reference which also has an adjustment as well and that can be directly fed into the input switch. So when you select the input, one of the scalers that you can select is the 8volt and that's going to inject the 8 volt directly into the comparator over here and that will allow you to calibrate as well. That's also pretty important because of course these RAM generators and everything else as we discussed are going to drift and the one way to establish the references is the zero and the 8VT. There's also many other adjustments inside of the unit controlling the ramp and other aspects of it. We don't have to necessarily play with those, but you can see why you would need these two adjustments because there's obviously so much drift that's happening. Another really important thing here is that remember when this was made, there was no ultra high input impedance devices and FETSS and so on.

[00:08:39]They had only diodes and maybe bipolars.

[00:08:41]So of course everything was leaky.

[00:08:43]Everything would consume current. Not to mention that every time they generated pulses, they would steal some current and some signal from the input, especially with the diode comparators.

[00:08:52]That means that achieving a really high input impedance would have been tough.

[00:08:56]Not to mention, as you steal current from the input impedance, you're changing the actual input characteristics of the circuit. So, what do they do to compensate for that? Well, they would actually inject charge back in, which is going to be the opposite of the charge that they're taking away. By doing that, they would null everything out and it would look like a ultra high impedance. This is pretty clever and really subtle and difficult to implement, but they had no choice because of the fact that every time they did anything, they consumed a little bit of current. So, by injecting some charge back, they emulated that ultra high impedance that you would need for the multimeter. Pretty clever. And the input section of this module is probably the simplest part of this entire system. You can see that the input signal coming over here goes through a long chain of voltage dividers. And that's going to just give us the the decade ranges that you can select between each of those signals then go into that switch 10vt 100vt and the cal 8volt input can be switched in directly. That's really all it is doing. There's also an RC filter in the front to attenuate everything above DC. I think 30 dB of attenuation at 60 Hz or so because this thing can only measure DC voltage. And then all the complexity is sitting on the right side of this. Some of the stuff we talked about. I think we've probably looked at this enough for us to go ahead and play with it. look and see how it is constructed and then do some measurements. See how close it is to being a good multimeter after all these years. And by the way, I've been using the word multimeter and voltmeter interchangeably here, but they're of course not the same thing at all. In reality, this is just a DC voltmeter and nothing else. Although, if you put a parallel resistor directly at the input terminal and pass current through it, you could use it as an ammeter if you know the value of that resistor. But that's more of a manual procedure. So, in reality, this is just simply a manual scale DC voltmeter. So, let's take a quick look inside of this module using the Tago T50 microscope, which has been working really well. By the way, if you've noticed in several of the previous videos, these really high quality PCB shots, they're all filmed on the Tago T50. And the board here at the bottom is where we have our input voltage divider in combination with the switch over there. That's going to give us the input range. And there's a couple of precision resistors over here.

[00:11:02]precision being 02% which is part of that long chain of voltage dividers we saw earlier and some potentiometers for adjustment and I don't think any of these have ever been adjusted so this might be brand new and then we do have the voltage ultimately feeding from this side to the other side of this other board which is going to be our input comparator this is probably the ground comparator and that might be the ramp generator with various individual adjustments that can be done on that as well there's some capacitors and everything else in there some input capacitor there as well and over here we can see our input switch. There it is.

[00:11:32]You can turn that to do the various selection. And there's two potentiometers in there worth paying attention to. So this potentiometer is our zero adjustment. And the one underneath it, if you look carefully, there are three ball bearings buried down in there. This is a really interesting design. So those ball bearings make up a planetary gear system. So which means that you can then adjust that potentiometer in a much much finer way with the same one turn potentiometer. So when you turn the input, they're going to go through the planetary gear. They have a huge reduction. So you can turn the input many, many times. Even though the potentiometer itself is only a one turn pop. It's just a gear reduction, which is a lot of trouble to go through. But I think just back then they didn't have very good potentiometers to be able to do this. Yeah. Other than that, nothing unusual going on. Some really, really old components in there. Looks actually really beautiful under the microscope as to be expected. But yeah, I'm really eager to try it out. I don't want to make any modifications to it. I just want to know how it works after all of these years of sitting in storage. And now that we have some good idea about how this digital voltmeter plugin actually works in this frequency counter, we can give it a try. You can see that it's already installed into the chassis itself. And we just have to now configure these knobs to take advantage of the plug-in module. So this goes into the plug-in. The time base is on 0.1.

[00:12:50]And here we have the remote enabled.

[00:12:52]That's it. We just now have to turn it on. And we should be measuring volts.

[00:12:56]You can see that this is set to a 10vt scale which is the lowest scale we have available and let's wait for it to stabilize a little bit as well as of course come up to temperature before we do any zeroing or any calibration. Okay, we have warmed up a little bit and everything is now stable. Now we should try and zero the instrument. You can see that we are indeed reading a little bit of an offset over here. Remember we're on a 10V scale which is the smallest and we are reading 0.21 volts or so which is actually not that large. Now remember in order to zero this it's best to short the inputs first. Let's go ahead and do that.

[00:13:27]There's the shorter the input. You can see that the number changed by just a little bit. Now, what does zeroing this situation even mean? Since we are reading a negative number, when you zero this, you basically want to half the time read a positive and half the time read a negative, meaning that it's oscillating around zero, which means an average of 0 volts. That's the ideal situation. So, let's go and adjust that.

[00:13:46]See if we can bring it to zero. If it doesn't go to zero, that would be a problem. And let's see. Oh, there we go.

[00:13:51]We're getting better. And look at that.

[00:13:54]We're still at not oscillating. There it is. Look at that. That's pretty good.

[00:13:58]That's almost perfectly 50% duty cycle between the two. That's good. So, now let's see what the 8VT actually looks like. I don't need to remove this because it's physically disconnecting it. And here's the 8VT. Actually, that's not so bad neither. Let's see if we can adjust that and bring that to an 8 volt.

[00:14:14]This is the potentiometer with the planetary gears on it. Let's keep going.

[00:14:19]Oh, it is working. Oh, that was too much. Whoa. Oh, it's actually really sensitive even with that.

[00:14:26]I think the potentiometer might be a little bit dirty. Yeah, I think maybe if I just turn it a little bit back and forth might actually help just so we run those carbon fibers a little bit back and forth.

[00:14:38]Okay, not so bad. There we go. And I should be able to hit almost perfectly 8 volts. Look at that.

[00:14:46]That's not so bad at all. There you go.

[00:14:49]I think I'm reasonably happy with that.

[00:14:51]But of course, this test doesn't tell us anything about the absolute accuracy of this instrument because we're only validating it against its own internal reference. In order to truly test it, we have to check it against the real calibrator. So, let's do that next. So, now let's use some serious calibration instruments to validate the performance of our frequency counter with a voltmeter plugin. At the bottom, I have the Fluke 5728. We're going to generate some voltages using this all the way up to a,000 volt, which is full scale for that, and measure that. And over here we have the Fluke 5798. This is just to measure this voltage to show us that it is actually working. So we start with one volt that only has a 10 volt as its minimum input range there. So at 1 volt we're going to enable that and we should see something really close to 1 volt at the top there. Give it a second to stabilize. There it is. Okay. So that's pretty good. So we should be able to take that 1 volt and put it directly over here. You can see that the 8 volt is being measured really well. So that looks good. Let's go ahead and enable remove this short over here. Let's disable that. Take this one from here.

[00:15:49]Plug it in over here and enable. Let's see what we read as one volt. I got to bring this all the way down here. And look at that. Not so bad. It's only off by a little bit. I would say that's pretty accurate after all these years.

[00:16:02]It's also pretty precise. So, let's go ahead and try something higher. Let's try 5 volt over here.

[00:16:08]What do we get? Okay, so it's off by about maybe 2.8 molt or so. Honestly, that's not so bad.

[00:16:14]What about 10 volt on full scale over here? And there's a 10vt. Pretty good.

[00:16:19]Off by a little bit as well. Looks like there's a linear offset there. So, here's a 100vt input. Yeah, looks pretty good as well. So, we can now set this to 50 volt.

[00:16:30]And have to enable that again. And there's our 50 volt. Honestly, not so bad considering how many years this has been sitting around. Let's try another 100 volts over here. There is our 100 volt. Yeah, you can see that the offset is almost exactly the same in all the ranges, which kind of makes sense. So, that looks pretty good. Should be easy to calibrate if you really wanted to.

[00:16:50]And here's a,000 volt input. Here's 100.

[00:16:53]Let's do 500 volts over here.

[00:16:56]And there's our 500 volts. Looking pretty good. And ultimately, here's 1,000.

[00:17:04]Look at that. 1,000 volt. So, it's reading.7 volt off, which is again quite a bit. We can also validate that onto this instrument as well to make sure that it is indeed correct. Let's put this into the 1 kilovolt range. Take that from here.

[00:17:20]Plug it over here. Double check. Make sure this is in 1 kilovolt range. That is good. 1 volt and enable that. Let's see what we get at the top over there.

[00:17:30]Wait for it to stabilize. Yeah, there it is. 1,00.018.

[00:17:36]Again, this hasn't been calibrated nearly as well as this one has, but that's pretty good. It's considerably better than what the other instrument is showing. So yeah, that's pretty impressive considering how long this has been sitting around. I think it's doing pretty well. And there you have it. I hope you enjoyed this quick video about how this counter can be converted into a voltmeter and give you some more insight about how people back then thought about creating instruments like this. As always, supporting the channel goes a really long way. Patreon and PayPal are both available. The link is in the description. And I always put everything that comes back into the lab so we can make more interesting videos in the future. Subscribe. I have a lot of cool things coming up. And I'll see you in the comment section.