Budget multimeter recommendations?

[DanceswithUnix]

You can't just poke probes at an open PWM pin and expect to see something useful - You have to read the interface spec and meet the requirements for pull up / pull down and decoupling Etc.

There is a lot of glossing over of details going on for my part, but this is something that I meant to say and it did get omitted.

A decent modern motherboard will only output a PWM signal if it detects a PWM capable fan. I would try and grab a 4 pin extension cable to make measurements easier, but you probably need a fan plugged in to see anything. A circuit to fake there being a fan connected would cost more to make than cheap fan

In that circumstance, I would totally expect to be able to see the PWM signal even on one of those cheap multimeters. He isn't talking about detailed duty cycle measurement here, just a basic "is this actually PWM" which should be pretty obvious compared to a DC fan control with just the voltage varying.

I'm very much a learn by doing sort of person so I'm going to buy the VC99 you linked to earlier and the updated model of the DSO138, when it's in stock in the UK that is because not knowing if i had to pay import fees would trigger my anxiety, that way i can learn how to use a DMM and practice soldering while i build the DSO and not have to worry to much if i get something wrong.

That DMM says it can do frequency and duty cycle measurement, in which case it should be able to tell you what's going on, or at least I expect I could work out what it is doing with that but I'm not sure how clear it would be to a beginner. It looks a useful bit of kit regardless.

That diy scope still looks rather toy like. Having some sort of 'scope is really nice though, nothing beats being able to actually see the waveforms coming out, specially if a circuit was designed by someone like me where I have used some pretty whacky waveforms for driving things like LEDs because I ran out of hardware timers and had to drive off a 1KHz interrupt in software, but it saved a few pence doing it that way which adds up when you sell a million units

If you want something that could be genuinely useful in the future, you sadly need to spend proper money, something like this:

https://uk.farnell.com/multicomp-pro...sps/dp/3107575

where a PC does the display and user input bits, so you are only paying for measurement electronics. at work we use similar Picoscope kit, I think we only have one "proper" scope with a display on it.

[matts-uk]

Sorry, i guess that's what happens when a luddite like me doesn't know what they're talking about.

No need to apologise, I doubt you are a luddite and the only thing to be embarrassed about is the expectation. An electronics degree is 5 years. I took a more vocational route but it was still 18 months of education and lab work to prepare for the work placement. I merely scratched the surface. I wasn't an expert then and I am not one now.

]It's very much appreciated that you've taken the time to write a proper explanation as I'll undoubtedly come back and re-read it when the doing/practice makes something click in my brain, probably doesn't help that maths doesn't come naturally to me and even things like Ohm's law is just a word to me ATM.

Hmm. Yes. I hear stuff like that every time I take on a new technician. Generally they take longer getting over their adversity to doing maths than doing the maths. Poking around with electronics without a grasp of Ohm's law is like attempting to drive a car without first learning where the accelerator, clutch and brake are, and the relationship between them - I guess it's possible but it sounds painful and expensive to me

I may not even be able to workout what i set out to due to ITE not publishing the specs of the controllers so I'm guessing i won't know what the requirements for pull up / pull down and decoupling Etc are, I'm OK with that if I'm honest as it's become about learning more about the things all you guys have talked about.

It's easier to learn about 'the things' from first principles. Start with simple circuits and build on that. You can have a lot of fun with an Arduino and a handful of resistors and LEDs at pocket money prices. I mean there's a bit of maths but it doesn't have to be boring.

Take the time to do the background work and you don't need the ITE data-sheet either. I can point you at the materials but it's up to you to read them.

Step 1. How is it supposed to work and why?
https://www.analog.com/en/analog-dia...an-speed.html#

Step 2. Find similar schematics
Scroll down to Figure 7 in the Analogue Devices document.

The motherboard will look similar to this.

Ref: https://www.electroschematics.com/4-wire-pc-fan/
There is a salutary tale in the second link

That's why i prefered the DIY kit thing as it will give me a chance to learn as i go and due to it being partially open source and cheap i can try/confirm things without the worry of turning £50+ into blue smoke.

So buying advice.

The VC99 is made by Vichy and sold under many different names. Typical Chinese OEM stuff, looks better on the outside than it is on the inside. It will do the job but don't expect it to last. I wouldn't be brave enough to test the safety rating by poking at a mains supply though. Personally I wouldn't bother with the Vichy. I would put the money towards a known brand with a reputation to worry about. For a beginner, I'm thinking Uni-T. Another Chinese manufacturer but they have been around since 1988 and include a cal cert in the box. Uni-T are like the Draper tools of measurement instruments. You get a lot for your money and the quality is not that bad. If you don't mind a discontinued model the UT61D is a legend of a hobbyist DMM. Comprehensive package of useful features, with RS232 output and freely downloadable software for data-logging and scope-a-like charting. About £50 on e-bay from Lithuania or China.

The pocket scopes get a bashing, quite deservedly, but are not entirely useless. I have one I use for the odd sanity check. Easier than pulling out my full size DSO, less risk of electrocution and at £20 I am not going to cry if I blow it up. It helps to know the limitations as those limitations can make a perfectly fine signal appear screwed up. You would be better buying a ready assembled example IMHO. If you build it and it doesn't work when you switch it on, what are you going to use to fault find? Soldering practice is best done with some strip board and a handful of resistors and LEDs.

[ik9000]

I've tested the mains with my VC99 (and a plug probe converter). Worked fine. I've tested numerous guitar+amp+pedal+psu electronics and it's been fine. It's not a fluke and I wouldn't use it in a lab/for military contracts etc but it is fine for the amateur hobbyist as a starting point. I wouldn't push the OP to splash more money than they need to when they clearly are just putting their toe in the water so to speak.

@OP a great textbook if you have a basic physics background under your belt already (and tbh even if you haven't) is Practical Electronics for Inventors by Paul Scherz and Simon Monk (Mcgraw Hill). I have the 3rd edition in print and 4th ed on kindle. 4th is much the same but some comments about thinner paper in the print edition. It goes from the basics very well, and better than my uni lecturer did tbh!

Now on the fans themselves. Had a spare 5 minutes last night:

I tried seeing what the vc99 would come up with measuring the tach and pwm pins. Akasa fans tach measured the fan rpm (albeit in 2xHz) (46Hz reading is 2x23Hz as presumably double pulse per revolution due to the split coil) but Noctua fans same tehnique spat out 2kHz so no idea what's going on there, and indeed whether the akasa fan was a fluke or whether the Noctua uses a different system (and therefore how the same header can control the two different types so readily), (both were pwm fans).

I need more time to play with the pwm as I only played with 100% which obviously won't be doing any pulsing. I agree with others' comments it's probably not the best way to try and measure fan behaviour, and hence I also use that Lamptron board I mentioned to do my checking of what's going on, but it should be able to detect something from the pwm pin. Whether that is of any use remains to be seen!

[matts-uk]

A decent modern motherboard will only output a PWM signal if it detects a PWM capable fan. I would try and grab a 4 pin extension cable to make measurements easier, but you probably need a fan plugged in to see anything. A circuit to fake there being a fan connected would cost more to make than cheap fan

'Detect' sounds too grand. The PWM signal drives a FET. Without some bias on the line you are going to see jack all. Faking the PWM input can be done with a couple resistors. Let's be generous and call it £1.

The learning point is the meter needs to be in parallel. A breakout is ideal. Alternatively, the PCB header is usually long enough to get a hook or bent wire under the connector but care is needed to avoid a short.

In that circumstance, I would totally expect to be able to see the PWM signal even on one of those cheap multimeters. He isn't talking about detailed duty cycle measurement here, just a basic "is this actually PWM" which should be pretty obvious compared to a DC fan control with just the voltage varying.

For a functional test you need any old meter that will read +12V DC - My car battery charger would do.

If the fan runs at less than full speed with +12V on the Yellow wire, it's PWM.

That DMM says it can do frequency and duty cycle measurement, in which case it should be able to tell you what's going on, or at least I expect I could work out what it is doing with that but I'm not sure how clear i would be to a beginner. It looks a useful bit of kit regardless.

I'm still smirking at a 30Mhz frequency range on a 4000 count no-name Chinese meter - More aspiration than specification

PWM frequency measurement can be misleading. As I said, 100% duty cycle will read 0Hz and the supply voltage, exactly the same as a dumb fan. Maybe the reading isn't 0 or the PWM frequency, is that a fault? The unexpected reading can be caused by a DMM not being suited to frequency counting - The lack of a trigger control gives it away. The % (duty cycle) key can be misleading too. My understanding of the spec is the PWM signal is sometimes inverted (active low).

Owning test equipment is not enough. You need to know what you are measuring and how the test can influence the result.

I have used some pretty whacky waveforms for driving things like LEDs because I ran out of hardware timers and had to drive off a 1KHz interrupt in software, but it saved a few pence doing it that way which adds up when you sell a million units

Did you plonk an LED on the end of the interrupt pin and expect there to be enough current? I'm guessing not.

If you want something that could be genuinely useful in the future, you sadly need to spend proper money, something like this:

In my experience owning a similar pocket scope, a surprisingly useful toy for £20. I bought mine as a stop gap when my old CRT bit the dust. The original firmware needed binning but after flashing BenF-Waves it's been OK. You wouldn't want to measure anything with it but as a glorified logic probe, yeah, it's cute.

...at work we use similar Picoscope kit, I think we only have one "proper" scope with a display on it.

If you only had one scope, which would it be, PC or stand-alone?

[oldboy47]

Gentlemen, from me, thank you.

Thank you to Corky34 for asking the questions and Matt, Ik, Cat, Dances et al for answering without being condescending or biased.

For me, it is why I like Hexus

[Corky34]

...but Noctua fans same tehnique spat out 2kHz so no idea what's going on there...

Funny that you should mention Noctua as their new A12x25 PWMs are what started all this, well that and Gigabyte motherboards, apparently they don't play well together which kind of sucks when you discover that a year after buying 4 of them.

As it doesn't seem to be something Gigabyte is ever going to fix and i can't update the fans i thought, maybe foolishly, that if i learnt about electronics there's a remote chance of doing something about the fans stopping when going under 1000 x rpm and if not I'll at least have a chance to learn something along the way. Also a coincidence that matts-uk has mentioned the Arduino as in all the reading I've been doing I've seen it mentioned a few times so thought I'd read up on that before bed.

[CAT-THE-FIFTH]

Funny that you should mention Noctua as their new A12x25 PWMs are what started all this, well that and Gigabyte motherboards, apparently they don't play well together which kind of sucks when you discover that a year after buying 4 of them.

As it doesn't seem to be something Gigabyte is ever going to fix and i can't update the fans i thought, maybe foolishly, that if i learnt about electronics there's a remote chance of doing something about the fans stopping when going under 1000 x rpm and if not I'll at least have a chance to learn something along the way. Also a coincidence that matts-uk has mentioned the Arduino as in all the reading I've been doing I've seen it mentioned a few times so thought I'd read up on that before bed.

I bought two of the slimline ones and had PWM issues with my Asus motherboard trying to run two off a splitter. Both would work,and then randomly one would stop. I was not sure whether the splitter was the issue,but I am sure they worked with some other fans I had. Must check and see if the single one I have is working correctly too.

[ik9000]

Funny that you should mention Noctua as their new A12x25 PWMs are what started all this, well that and Gigabyte motherboards, apparently they don't play well together which kind of sucks when you discover that a year after buying 4 of them.

As it doesn't seem to be something Gigabyte is ever going to fix and i can't update the fans i thought, maybe foolishly, that if i learnt about electronics there's a remote chance of doing something about the fans stopping when going under 1000 x rpm and if not I'll at least have a chance to learn something along the way. Also a coincidence that matts-uk has mentioned the Arduino as in all the reading I've been doing I've seen it mentioned a few times so thought I'd read up on that before bed.

Wait, so what is the issue? Your fan is stopping, or you want to make it stop?

From a previous post you suspect the mobo has fake pwm 4-pin header that you think isn't doing pwm. That is an easy check with a Y cable and a multimeter.
1)1 buy some of these:
https://www.amazon.co.uk/Elegoo-120p.../dp/B01EV70C78
2)peel of some of the male-female ones and you can plug those to the fan Y cable header, and use crocodile clips onto the male end from your multimeter set to DC volts.
3) connect the two left-most pins (yellow and black cables) and see what happens to the DC voltage as the fan speed is adjusted. For a pwm fan the voltage should be constant (12V) across those two pins. If not then it's changing the fan voltage via conventional 3-pin method.

Now if the fan is cutting out below 1000rpm (how are you confirming the rpm btw?) That fan according to Noctua has a min speed of 450rpm. If it is receiving pwm duty below what it needs for 450rpm it will stop. If you want to make it stop below 1000rpm then you can configure it in the mobo uefi usually or using fan control software along with setting the temp you want the fan to spin up at. I wouldn't do that for a cpu fan but for certain case fans sure I can see a case (if you pardon the pun).

How many fans are you using from the header? Just one at a time or are you using the fan on a Y-cable or larger splitter? What current rating does the header have?
The noctua nf-a12x25-pwm only has 0.14A draw so well within most headers. Which other fans are on the splitter? Do they all have the same current draw?

I have played with the 140mm Noctua NF-A15 PWM and NF-A14 PWM as well as the Chromax NF-A15 HS PWM and they all respond to voltage control. At 100% they are 1620rpm, at 40% pwm they are at 960rpm. My pwm controller limits it there and jumps to 0% to prevent weird stuff. However muggins here hooked them up to a voltage controller for giggles. Here's how they behave using 100% pwm and changing the voltage supply:

12V 1620 rpm
11.5V 1560rpm
11V 1500rpm
10V 1380
9V 1260
8V 1140
7V 1020
6V 900 and the limit of my little asetek fan controller. FWIW 600rpm is practically silent - I mean stick your ear next to it barely a sound, and inside a case you're not hearing it.
I now stick in some resistors (totalling 45ohm in series)
5.5V 820rpm
5V 720
4.6V 660
at this point I think, ok throw in some more resistance.

Swap out the previous ones, add in 200ohms in series the fan wouldn't turn on. Not enough juice to get itself spinning. cranking the asetek dial back to max saw the fan grumble and kind of try but it was not happy and I just took the 2nd resistor out.

Now with 100ohm in series (roughly the same as the fan's internal resistance)
weird stuff happens. The lamptron logs the fan going right down to 240rpm (and visually it is definitely spinning a lot slower) but the multimeter voltage reading is jumping all over the place from 4.6 to 6V constantly, not settling at all. The multimeter is not the issue - it can definitely handle that low voltage range, and with the fan being slower the supply voltage should now be well below the range it keeps reading as. It's almost like the fan has suddenly started doing its own pwm! Either the fan is now slow enough for the coil switching to be giving some weird ac effects back to the multimeter (I really don't think it should be) or alternatively the asetek controller is fighting the current sag, but then I'd expect the fan speed not to reduce like it does. The supply current at source is rising but only from 0.3A to 0.32A so I'd be surprised if it's that. I'll see if I can rig up a separate 12V supply to run the lamptron display and just use my power source to drop the supply V to the fan right down in isolation and take the asetek controller out of the loop. Having worked out the circuit for the asetek it's not a simple voltage divider, it's using a voltage regulator transistor so could be interfering. I'll see if the DMM plays better in AC rms mode.

TLDR? The 140mm version will definitely go down well below 1000rpm however.

The point I'm trying to get to is if your UEFI is like mine you can set that fan in the BIOS to be a traditional (non PWM) fan and run it that way. It should behave just fine going by my tests on its larger sibling.

I bought two of the slimline ones and had PWM issues with my Asus motherboard trying to run two off a splitter. Both would work,and then randomly one would stop. I was not sure whether the splitter was the issue,but I am sure they worked with some other fans I had. Must check and see if the single one I have is working correctly too.

If an old board it's possible that might be a current supply thing though most headers can handle at least two fans these days, usually three or four.

[CAT-THE-FIFTH]

If an old board it's possible that might be a current supply thing though most headers can handle at least two fans these days, usually three or four.

Its an Asus Strix B450I mini-ITX motherboard.

[ik9000]

Its an Asus Strix B450I mini-ITX motherboard.

where are the fans in the case cat? Is one of them getting interference from another fan, say the cpu fan or similar that might be helping one keep going while the other struggles?

[Corky34]

Wait, so what is the issue? Your fan is stopping, or you want to make it stop?

The point I'm trying to get to is if your UEFI is like mine you can set that fan in the BIOS to be a traditional (non PWM) fan and run it that way. It should behave just fine going by my tests on its larger sibling.

It's that they stop, well sort of, even when no stop is set in the BIOS. Between 25(ish)% and 35(ish)% PWM set in the BIOS they'll spin on first boot then a reboot, sleep or just falling bellow 1000rpm for to long and they'll stop spinning and go into a "I'm about to spin", "No I'm not", sort of jitter. Like they're getting a kick of voltage but only 1 kick per second. The slowest they go when the BIOS is set to voltage is around 850rpm so that's what I'm using ATM.

They do that when connected directly to a header (1x1) or when there's two running through a splitter, it's also worth noting that this only happens with their new A12x25, the ones with more blades that look a bit like gentle typhoon's.

The rpm is being taken from what the motherboard control chip is reporting, well when it's not jumping around from reporting 0rpm to 1000rpm to 4000rpm every second.

I assume the headers have the standard 1amp rating but Gigabyte being Gigabyte (it's one of their X570's) they don't actual say what amperage they're capable of, either way I've tried them on both fan and fan/pump headers (that i assume are higher amperage), no difference.

RE: Your 1, 2, 3 trouble shooting steps that's sort of what i intended to do, at least something similar, basically watching what happens by taking measurements using a spare Y splitter.

This isn't the only reason i wanted a DMM and a play around with a DIY scope BTW, i maybe stupid when it comes to electronics but I'm not crazy, I'm just getting there is all.
Since fixing an old monitor a while back I've always wanted one and to understand electronics/electrics more than my typical sweat inducing that shouldn't be live because it's switched off or I've pulled the breaker but I'm not entirely sure, lets touch it and find out approach.

[DanceswithUnix]

'Detect' sounds too grand.

Really? It can vary voltage and PWM and check the tacho pulse return to see what works, with some careful finesse to check whether PWM works before you attempt voltage control as voltage controlling a big 140mm or bigger fan will likely blow the motherboard header.

Did you plonk an LED on the end of the interrupt pin and expect there to be enough current? I'm guessing not.

I don't know what you are trying to get at there, but many microcontrollers are rated for direct LED drive (as long as you are sinking current, not sourcing). In that case though I think we were driving three LEDs from one pin, so we buffered it with a transistor.

[DanceswithUnix]

The rpm is being taken from what the motherboard control chip is reporting, well when it's not jumping around from reporting 0rpm to 1000rpm to 4000rpm every second.

That doesn't sound good

I'm glad you have said what the actual problem is though, it should make it somewhat easier to help. The thing is, 4 pin PWM headers are cheaper for the motherboard, so there isn't any reason to "fake" it. OFC that doesn't mean someone hasn't faked it for some bizarre reason, but it seems more likely to me that they simply messed up with their fan control and what you are looking at is just a bug in their control system.

Is it possible to set the BIOS to forced PWM rather than voltage control, and set a fixed fan speed which is fast enough to keep the CPU cool without being deafening? I'm not saying to use that as a long term solution, just wondering if avoiding all the fancy PID fan control logic and driving it old skool would give a stable fan speed.

[ik9000]

voltage controlling a big 140mm or bigger fan will likely blow the motherboard header.

Sorry but that doesn't make sense to me. Voltage controlling that 140mm fan last night pulled 0.3A total from the psu, and most of that is driving the lamptron. The fan only needs 0.13A at 12V. Given most headers are 1A these days it sounds like plenty of headroom to me.

It's that they stop, well sort of, even when no stop is set in the BIOS. Between 25(ish)% and 35(ish)% PWM set in the BIOS they'll spin on first boot then a reboot, sleep or just falling bellow 1000rpm for to long and they'll stop spinning and go into a "I'm about to spin", "No I'm not", sort of jitter.

That behaviour is normal for low duty% pwm on most fans tbh. A common recommendation is to set BIOS to not use <40% pwm and it should stop that problem, but you won't get the fan speed lower than the threshold rpm that corresponds to. To do that you then need to add a voltage drop to the 12V feed to lower it instead and maintain the cycle at 40% (or possibly raise it depending on what you encounter).

Measure the pin voltages on the board. Is ground true 0V and is 12V actually supplying 12V? That is first thing I would be checking.

Which gigabyte x570 are you on? Master/Elite/Pro etc etc and which header are you using out of interest?

[ik9000]

That doesn't sound good

I'm glad you have said what the actual problem is though, it should make it somewhat easier to help. The thing is, 4 pin PWM headers are cheaper for the motherboard, so there isn't any reason to "fake" it. OFC that doesn't mean someone hasn't faked it for some bizarre reason, but it seems more likely to me that they simply messed up with their fan control and what you are looking at is just a bug in their control system.

Is it possible to set the BIOS to forced PWM rather than voltage control, and set a fixed fan speed which is fast enough to keep the CPU cool without being deafening? I'm not saying to use that as a long term solution, just wondering if avoiding all the fancy PID fan control logic and driving it old skool would give a stable fan speed.

If the OP can set min pwm% and just wants to reduce max rpm noise then they can use the bundled Noctua noise reducing cable, which SFAIK basically shoves a resistor in series to drop the voltage feed like I'm suggesting. Personally I want something hybrid that allows the user to both vary voltage and also set a PWM offset like their NF-FC1 controller, but where you can select a curve profile and combinations of the two across the range. Granted that is more complicated, but it is achievable. Whether enough people would buy it at whatever price it would end up being is another matter...

[matts-uk]

[QUOTE=ik9000;4282599]

I've tested the mains with my VC99 (and a plug probe converter). Worked fine. I've tested numerous guitar+amp+pedal+psu electronics and it's been fine. It's not a fluke and I wouldn't use it in a lab/for military contracts etc but it is fine for the amateur hobbyist as a starting point.

The VC99/99+ has been pulled apart many times and it's cost-cutting design issues are well known. Probe sockets soldered directly on-top the PCB, untidy hand soldering, glass fuses instead of ceramic, insulation tape instead of physical distance, not a PTC or MOV to be seen. The meter does perform to spec out of the box. The trim pots aren't lacquered and it might not remain in spec for very long. It will work fine, until it doesn't. It is what it is, a 'cheap' Chinese OEM meter. You do as you please but sorry no, *I* am not poking something like that at 240V AC.

There are a lot of competing DMMs between a VC99 and even the least expensive Fluke. Uni-Trend don't compete with Fluke, any more than Draper competes with Snap-On. Uni-T meters do show up in all sorts of different scenarios and industries though, because the brand has been churning out 'inexpensive' budget instruments for 30 years. Even the contestants in the Pottery Throwdown were sporting Uni-T temperature meters on Sunday. They are like a red rash that spreads anywhere there is a need to take measurements electronically

I wouldn't push the OP to splash more money than they need to when they clearly are just putting their toe in the water so to speak.

If I could buy a much better designed and built product for a little extra cash, I would want to know about it. The VC99 is what, £35 delivered? The UT61D retails around £50 but £43 is common and if you look hard enough it's out there for £30 at the moment with free delivery.

The UT61D has probably seen more cash strapped EE students through college than any other meter in history. There are loads of tear-downs on Youtube if you want to compare the insides with the Vichy. No it is not a Fluke and no I wouldn't trust the 1KV safety rating and yes *I* would pay (significantly) more for some additional resolution. A first meter for a hobbyist, occasionally poking at 250VAC, I honestly can't think of a better value sub £50 meter.

Noctua fans same tehnique spat out 2kHz so no idea what's going on there, and indeed whether the akasa fan was a fluke or whether the Noctua uses a different system

Noctua have very helpfully published their specs. According to Noctua, a 2KHz signal would equate to 60,000RPM(!) If you had the correct range selected, and didn't misread a decimal point, it's likely you measured noise - No trigger control remember. Another possibility is you shorted the 12V line while prodding about and entered 'broken' undefined territory.

I agree with others' comments it's probably not the best way to try and measure fan behaviour, and hence I also use that Lamptron board I mentioned..

And you thought to pull me up on splashing cash? I take it you have the Lamptron for reasons other than measuring fan speed.