need help with an anlogue 360-440Hz 200v frequency meter

Here's one suggestion for a self-powered meter circuit. It full-wave rectifies the incoming AC and uses a zener to derive a supply voltage Vsupply in the range 9V-12VDC. The AC waveform also clocks a CD4013 latch to give a 50% duty cycle square wave at half the incoming frequency. The latch Q output triggers a monostable (U1a, U1b) to give pulses set by R4 to have a width equal to the period of the minimum frequency of interest (assumed 360Hz). Logic U1c, U1d combines the latch and monostable signals to give output pulses of width proportional to the difference between the incoming frequency and the set minimum frequency. These pulses are fed to the meter and are averaged by the meter's inertia (a 100uF cap could be added across the meter if desired). R6 sets the current to 100uA at the maximum frequency of interest (assumed 440Hz).
Simulation shows that with a stable Vsupply the meter drift is ~ +- 1% over the temperature range 10C-30C, but with a 5% variation in Vsupply the drift is ~ 6%. The zener could be replaced by a more accurate reference (LT1021-10, LM317?) to reduce drift.

Hi here is a picture of the meter and equipment im trying to calibrate it with. Alec are u familiar with Tina industrial , i drew your circuit up using Tina and came across some problems mb my fault i dont know but i couldnt get it to work as you say. I think i am having difficulty getting the input to the circuit right also you have not stated what voltage the capacitors are and what the 5 diodes are they could be anything. I tried drawing the circuit using LtSpice which i think is what you used but the two chips were not in the library and when i tried to add them it went wrong. I have decided to build the circuit you have suggested a mate of mine is good at etching boards but i need to know the voltages of caps and what type of diodes u used . I can draw basic stuff using Tina but when it comes to frequency stuff im ****. If you or anyone here is familiar with Tina can someone explain to me how to set a 200v 360hz input to the circuit using a voltage source, when i open up properties of voltage generator all i see is DC level (V) which i set to 200 then underneath i set it to sine open box nest to sine wave and that takes me to signal editor and i see Amplitude [v](A) now do i set this to 200 then underneath it says frequency which i presume i set to 360Hz then under that it says Phase[deg](P) which im not sure about so i leave at 0 , are these settings right or do i have to use the signal compiler which is second box from right side at top of the editor?? This is where i think its going wrong for me.Also where u have put Vsupply next to 470r resistor and the cap its a dead end which looks like a separate power supply along with a ground symbol, now that to me looks like 4 inputs , you said this was self powered, if it has a 200v A/c 360hz signal going in and a Vsupply dead end where is the power coming from for that along with the ground connection which a/c doesnt have just live and neutral. Please help as i am looking to buy components to have a go at building this board but i need to get a few things right so i can explain to my mate who is going to build the board for me, he know s some electronics but like me frequency stuff is just too much too grasp. Hellllllpppppp??!!! Any ways here is the picture of kit and instrument i promised if you look closely you can see machine output at 439hz on the thurlby and meter is wayy off , that instrument was calibrated on friday afternoon and it was bang on the end of scale and 360 was on 360 point on dial now look at it ,thats what i mean by unstable. If you look closely to bottom left of picture u will see the toroid that boosts the 50v out-put of machine to 200v needed for meter, this kit was used 10+ years ago and has only been stored i n a cupboard so kit should really be still ok. Let me know what you think guys. Still need help this is 2 weeks late for delivery and its getting close to giving customer £230 back and i dont want to do that unless i really have to.I just dont understand whats going on i have had a suggestion from a guy that component manufacture has changed so radically that this board should have been redesigned b4 i even said yes to the job. Another idea is if i could pay for the components someone could make them fit on the smallest board possible for me but i dont know which way to go.Let me know.

it's a signal generator? i think you should run your frequency meter off the 50V side, where you have more available current. signal generators don't produce much current. the frequency meter draws about 15-20mA in operation, that's why you're losing most of your voltage when you connect it. i thought you were running it off 400hz POWER, which is what it's made for. every time you adjust the pots, you change the amount of current the meter is drawing. it turns into a comedy of error sources.

Alec are u familiar with Tina industrial

Click to expand...

Sorry, no. So can't help convincingly re your Tina setup; but it sounds as though your settings are right for the 200V amplitude 360Hz signal. The 2 chip models I used in LTSpice are in CD4000.lib downloadable from
**broken link removed**
The lib file needs to be in the LTSpiceIV/lib/sub folder for it to work, and the schematic has to have on it the directive '.inc CD4000.lib'.

you have not stated what voltage the capacitors are and what the 5 diodes are they could be anything.

Click to expand...

The zener diode is 10V nominal and that defines the operating voltage of the whole circuit to the right. So 25V rating for the caps would be ok. All diodes apart from the zener are general purpose signal diodes (as you say, they could be anything...within reason), e.g. 1N4148. They only have to withstand ~ 12V reverse voltage.

Also where u have put Vsupply next to 470r resistor and the cap its a dead end which looks like a separate power supply

Click to expand...

Vin is the power supply. It is the 200VAC input you specified. This is dropped down by R1, rectified by D1-D4, limited by the zener to 10V, and smoothed by R2/C1 to provide a fairly stable 10V supply (Vsupply) for powering the rest of the circuit. But bear in mind what I said about replacing the zener for greater stability.

boosts the 50v out-put of machine to 200v needed for meter

Click to expand...

The meter circuit I posted will run happily with an input of anything from about 12V to 200V provided R1 is varied accordingly. So you don't actually need 200V. For a 50VAC input make R1 = 5k6. The circuit only draws an average current of about 5mA (and most of that is for the zener). Not sure where the 15-20mA figure in post #23 comes from?

Make sure the two ICs are standard CD4000 series, not high speed types (which can't cope with a 10V supply).
Use small presets for the two pots (trimmers). You should be able to fit all the components comfortably on a circuit board about 1.5 x 1.5 inches. If you use surface mount components then the board could be even smaller.

Good luck with the build. Keep us posted.

thanks alec your being most helpful, a few more questions tho the cd4093b has 2 connections not shown on your drawing Vss and Vdd i presume Vss is the connection to Vsupply and the ground line you have shown is Vdd but you have not indicated this on your circuit diagram, also cd4013B has a Vss and a Vdd pin i presume these are also linked to the same 2 pins as cd4093 and your version of CD4013 shows PRE which on the downloaded pdf version of the cd4013b is not on the pin-out, also is CLR the same as the reset pin on the same pdf ,D=Data and CLK=clock1 . So confusing bit is still PRE which you have put as ground ,but on pdf i think Ground is Vdd can u clarify these few points, and you also show a connection to CLK pin via a 10K resistor is this the power?? Sorry for so many questions but its getting confusing a pdf from FairChild semiconductors shows 1 thing and you have drawn many others. I have never used these so can u see where the confusion starts. Just so you know i spoke to an electronics supplier today whos very knowledgeable as a supplier goes and he says in 18 months time the CD4000 series are being phased out with no equivalents being hinted at, so buy em while you can guys cos they wont be around much longer, if i get this job right and i get another enquiry years down the line this will have to be looked at all over again.Let me know Alec. P.s unclejed did you not see the big toroid wired up to left of machine boosting 50vac to 200vac and higher which also boosts current too. And this equipment has been used before to do this job ,its just that in 10+ years since last used, components have changed properties enough to make this no longer work correctly, if i can get alec's circuit working and shrink it onto a small enough board i shall be using this idea from now on. One last thing Alec if i use the LM317 which i am also not that familiar with do i put it in place of the zener with the in side of it towards the 470r resistor line and the out side going towards the D4 line and adjust going to same line and no extra components?? to avoid further confusion could u show me how the Lm317 goes in place of zener, thanks,colin.

a transformer that steps up voltage does so with less current, if your turns ratio is 1:4 and your input is 50V at 40mA, then your output will be 200V at 10mA (actually, that's with a coupling factor of 1 and no other losses, and no physical transformer is perfect, so K will always be less than 1).

there's no such thing as a free lunch.

i presume Vss is the connection to Vsupply and the ground line you have shown is Vdd but you have not indicated this on your circuit diagram

Click to expand...

No. Vsupply goes to Vdd. Vss is the ground line. (They're not on the diagram because LTSpice takes them for granted).

also cd4013B has a Vss and a Vdd pin i presume these are also linked to the same 2 pins as cd4093

Click to expand...

Correct.

your version of CD4013 shows PRE which on the downloaded pdf version of the cd4013b is not on the pin-out, also is CLR the same as the reset pin on the same pdf ,D=Data and CLK=clock1 . So confusing bit is still PRE which you have put as ground ,but on pdf i think Ground is Vdd can u clarify these few points

Click to expand...

PRE = Set. CLR = Reset. D= Data. CLK = Clock. The 4013 chip contains two latches, only one of which is used in my circuit.

confusing bit is still PRE which you have put as ground

Click to expand...

It's grounded because it isn't used. Likewise CLR. Unused logic inputs should not be left floating, so the D, CLK, PRE (Set), CLR (Reset) of the unused latch in the IC should also be grounded.

you also show a connection to CLK pin via a 10K resistor is this the power?

Click to expand...

No. The latch is clocked by the rising voltage at D1/D2, but because that exceeds the chip supply voltage (Vsupply) the 10k resistor is included to limit current flowing into the clock input.

a pdf from FairChild semiconductors shows 1 thing and you have drawn many others

Click to expand...

Use the pdf data when building the circuit. Bear in mind that LTSpice only models components. The schematic I posted is what is used to simulate the circuit operation: it is not a complete constructional diagram. In particular, the schematic doesn't show pin numbers; you can get those from the pdf. Any problems, ask.
I'll post a pic of the mod needed for the LM317.

Here's the mod. I've only shown the left-hand part of the circuit. The right-hand part is as before.

Thanks Alec just a few more questions and i think im ready to have a go buying more parts and building 1 to test.
C1 i now see is polarity specific if its tantalum bead im in trouble cos i cant find one,but if its electrolytic metal can type im ok. Clarify please.
Also the C3 and C4 caps, i have found ceramic bead axial types are these correct? or do i need plastic type or ceramic disc type? Are all voltages of caps 25v or can they be slightly higher if i have trouble getting 25v. And the 39n one does not exist in RS can i use 33n or 47n . UPDATE RS does not have any small 47nF caps at the moment. BaHHHHHHH ,hoping 33n will work.Im sorry for asking these extra questions but i have an RS account where i get all my parts from.I understand all the info you have now given me, now when i build this and get it ready to go i guess i set 200v on rig and leave it alone just set the frequency adjust pots as necessary till it reads right on dial, i mb have to rescale dial from original scale in photo.Then when i remove instrument from rig and leave it for 30 mins it should read right when i reconnect. If 200 v has altered slightly from rig being left on if i readjust it back to 200v B4 reconnecting instrument it should all still read ok then it can be glued in case. finished off and sent to my customer . HOPEFULLY.!!!!!????? Many thanks for ur help Alec if i have any problems i will try be as accurate as i can diagnosing so mb you can help me further, with mb component changes etc etc. Let me know on this final info.

An ordinary aluminium electrolytic will do for C1. C3 and C4 are disc ceramic type. 25V or higher is fine.
If you use 33n instead of 39n you may have to increase the value of R4 above 100k (to set the zero point), e.g by connecting 10k (say) in series with R4.

Let us know if you have any probs. Glad to help.

**broken link removed**
hi me again i found these on ebay will they work for the 39nf cap in your circuit?
Also i just purchased 50- 0.1uf ceramic disc caps from ebay too as they seem to be the type u said as RS no longer stock these, will they soon start to be difficult to get hold of? Ebay has the ceramic disc type (39nf) but only from Hong Kong or Canada and i dont like the delivery times. Surely these will work or has it got something to do with the frequency they operate at? You did not say anything about my cal query so i presume what i said is correct about setting voltage to 200v B4 connecting meter then leaving it alone. Also will LM317LZ work RS. no.533-5711 , i wont need the version with the metal tab will I?

Those caps will be fine.

RS no longer stock these, will they soon start to be difficult to get hold of?

Click to expand...

I'm not surprised re RS, but I don't foresee any general shortage of disc ceramics. It's a sign of the times. Maplins even no longer stock resistors above 1 Megohm !

You did not say anything about my cal query so i presume what i said is correct about setting voltage to 200v B4 connecting meter then leaving it alone.

Click to expand...

If the customer is going to be using this meter on a 200V supply, use the 27k resistor at the input (R1) and use 200V when calibrating.

Also will LM317LZ work RS. no.533-5711 , i wont need the version with the metal tab will I?

Click to expand...

Should be fine. Metal tab version not needed as the chip dissipates so little power.

NOTE: Oops. I've just spotted an error in that circuit mod; there should be a 15V 0.5W zener (e.g. BZX84C15L) connected with its cathode going to the LM317 input and its anode going to ground. Sorry about that.

Edit: I've corrected the schematic in post #28. If this circuit is going to be in a confined space the electrolytic cap should be a 105 degree C type. I've also changed the values of R2 and R7 to reduce the power consumption. R1 will dissipate ~0.6W so needs to be at least 2W rating and should be well spaced from the rest of the circuit as it will get hot. Heatsink it if you can.

On looking through the LM317 datasheet again I realise the latest circuit may have a regulator load current less than the minimum required by the chip (typically 3.5mA but may be higher). This can be remedied by restoring R2 and R7 to their previous values, or trying 330 Ohms and 2k2 respectively as a compromise. Some experimentation may be required here and R1 will need to have a lower value (at the expense of greater heating) if the IC minimum load is near the limit of its spec (10mA). It may be worth considering an alternative voltage regulator IC. I'll look into this. In the meantime, I'd suggest put a hold on ordering parts.

Hi no parts for the regulator part of the circuit were ordered as i did not read the note on the bottom of the previous thread and i dont think the other parts were ordered i have a guy working on trying to place these components on the board this weekend and i cant contact him to hold off,, hes not going to be too pleased when i tell him on monday morning.Do u think the voltage regulator will still have 3 legs and the other things u mention will it just be a case of altering resistor values he wil not be pleased if more components have to be added as the board space is very tight as it needs to have the input connections in the middle of the board. Its going to be a double sided board as it is what with legs crossing over each other. Draw a 61mm diameter circle then cut it in half with 2 input terminals 35mm apart bang in the middle of the half board that is all the room we have to work with and even with surface mount chips its proving hard to cram it all on. Is there any way you could possibly have a go at drawing what it might look like as a completed board as i started to have a go at this this after noon and im running into difficulties , i got the right hand part with both chips and all connections up to meter connection side but i seem to have run out of room. What was the original values of R2 and R7 i didnt see them before you changed them. A 2W resistor is the largest i can get on the board and i cant heatsink it all i can do is mount it to board vertically not horizontally. Could you draw that 61mm dia circle and cut it in half and draw a line halfway from the cut line then find the middle as if you were making 2 equal quadrants then2 terminal points 17.5mm each side of that center line and see if you can get the components to fit ,thnx.

Hi i just found original resistors i did d/l and save the circuit they were R2=270 and R7=1K8. do i leave the cap mod for R1 as 4u7f or revert that back to 470uf as well? Do i also increase temp to 105dgc?

At the risk of incurring the wrath of your guy, I attach a reduced power version of the power-supply part. I hadn't appreciated just how tight a space you have to accommodate the circuit, so any power saving is an advantage. The good news is that it uses one less component (and possibly loses the heatsink) and shrinks two others.
Note:
R1 is now 56k and dissipates only 1/3 W, so a 1W type should cope without getting too warm and a heatsink is probably not needed.
C1 is reduced to 10uF.
R3 is increased to 22k.
D6 is now an 8.2V zener (anything from 8-12V will do).
The LM317 is replaced by a low-standby-current 5V reg (easily obtainable), e.g LT1021-5 or LM9036DT-5.0. These are available as 3-terminal devices or SOIC types, SMD.
R2 and R7 are replaced by C5 (10uF, 25V, 105C).

If you use all SMD parts I hope you can squeeze them all in. I'll try to find time to draw the circle and populate it

Attached is the semi-circle populated. Not sure if I've got the meter terminals in the right position.
Note: I used Eagle to draw this up and the component numbering differs from in my schematic. (I can post the Eagle version of the schematic if you like).
I found space for R8, which is a 2k PTC thermistor in series with the scale-max trimmer to help reduce meter drift with temperature. Simulation shows that it roughly halves the drift. Without R8 the drift is ~ 2% of FSD per 10C temperature change. This doesn't sound much but because the meter has an expanded frequency scale it amounts to 10% of the scale, ~ 8-10Hz per 10C change.
The meter is less sensitive to input voltage; a 10% input change gives a ~ 1% of FSD output change.
As the major source of heat is R1 (on post #36 schematic) and drift will be mainly due to temperature change, R1 should really be mounted outside the meter housing (perhaps inline with the input wire, or on a secondary external pcb).
I don't think any analogue circuit is going to have significantly less drift than this; a digital approach using a crystal-controlled oscillator for timing would be needed to get near-zero drift.

Edit: I forgot to mention that you won't find the CD4013 in the populated semi-circle. I removed that from the circuit (at the expense of a fraction of a percent accuracy). So that saves a small amount of real-estate. R3 (22k) then connects directly to the inputs of the first two gates in the CD4093.
I didn't use the Eagle auto-router so the component positions are probably non-optimum. I just wanted to see if the components could be squeezed into the space you specified.

Sorry alec no can do i need to go back to post #28 design for power supply , i cant find the 3 terminal version of the new regulator u propose and i cant justify £7.69 for cost of 1 chip, it takes board + meter over budget plus i have already got the 2 IC's (4093,4013) in surface mount version so i cant justify expense on 1 component im not going to use , they were purchased thurs. I have drawn up the circuit on paper with all the links for chips both above and below board i got nothing that crosses wires its just a case of trying to get it fit onto that bloody board. Sorry for the hassle Alec but i really do need you to try to fit original circuit plus post 28 power supply to that board. R1 cant be mounted externally but if its put near outer edge of semi-circle near to the flat edge it can warm up the inner edge of the aluminium case that will be glued onto base, just leave a little gap from other components i think it will be alright, the original board had a 27K 2w resistor on it that used to get warm when running it didnt affect anything in the past. I just wish i knew why the old circuit no longer works. Thanks for your time on this , let me know how it goes getting this post idea onto the board i reckon it will go i just cant draw it small enough to find out. I tried to get it to fit onto a board using Tina's pcb editor but im not that familiar with it and it looks huge and it damm fiddly.

https://uk.rs-online.com/web/p/linear-regulator/5165758/ will this 5v regulator work, its much cheaper 100ma max output max input 30v 5v output and it has 3 legs. I also cant go completely surface mount as a lot of parts are already in a box ready to be used over a few days last week they have been purchased and found i cant face much more expense here, running out of cash for this build and im gonna start losing money soon.

That would be fine except that it still needs at least 3mA, so the power dissipated in R1 is considerable. I can draw a layout using that, but positioning R1 is going to be a problem. What are its dimensions? What orientation and position? Can you post a sketch?
With R1 inside the meter casing I think the inevitable heating of the other components will cause unacceptable drift. If I'd known at the start that the whole circuit had to fit inside the meter I'd have suggested a different circuit
We could reduce the drift considerably and cope with a wide temperature range by using an NTC thermistor such as
https://uk.farnell.com/avx/nb12j00332jbb/thermistor-ntc-3-3k-0805/dp/1672283RL
(Alternatively a PTC thermistor could be used but they tend to be scarce and expensive). Can you source one in the range 2k to 4k7 ok? [Don't buy it yet!] That would enable the regulator to be eliminated altogether (NTC cost is about the same as 78L05 reg) and further space can be saved by using half-wave rectification to eliminate 3 diodes, but the input smoothing cap value would need to be increased.
I'll post the simplified circuit with these changes, hopefully later today.
I realise this messes you about, but designs inevitably evolve as difficulties and new facts come to light.