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

[alec_t]

i have already got the 2 IC's (4093,4013) in surface mount version

Is that the SOIC outline (pins 1/20 inch spacing)?

I also cant go completely surface mount as a lot of parts are already in a box ready to be used

Which parts? What are their dimensions?

 

[alec_t]

Ok. Here's the simpler circuit using a 3k3 NTC thermistor for temperature compensation. Theoretical drift is only 0.3% FSD over a temperature range of 0C-50C, so heat from R1 shouldn't be a problem (though it does dissipate nearly 1W).
I've had a stab at placing the components with Eagle, but it's guesswork until you can specify your known component dimensions. I've left a space on board for R5, which might be needed if a 3k3 thermistor can't be sourced but a 10k one can.

 

[eddi40]

Hi alec thnx again all resistors are RS 0.6w thru hole type, diodes are all thru hole type, caps are all thru hole type, only 2 parts surface mount are the 2 chips cd4093b and cd4013b which i must use both otherwise money wasted they have a length of 8.5mm and a width inc full leg length of 6mm( both chips have that diamension). I still need to use both chips.Also all resistors and diodes will be mounted vertically . Caps, chips and pots i have no choice with those. The 2 trimmers will be 1/4" square thru hole mounting side adjust multiturn type . I have had a go at putting all components on to the board using pcb design part of tina but i have run into problems, it will only show diode and resistor mounting as normal horizontal to board diamensions it wont just let me add pads to board how i want them. Using both chips as surface mount and others mounted vertically i can see this fitting i just cant draw it on Tina or on a piece of paper its just too fiddly. I cant work out on tina how to draw a semicircle to the right size. I have drawn the circuit out using the lm7 regulator on paper but i cant shrink it to the size of board on paper. Also those 2 holes in the board are the terminal pads which the board screws to using 2 2mm c/h screws then when meter works the 2 screws are soldered to prevent them coming unscrewed. I did say on post#34 2 terminal posts 17.mm each side of center line. Now i can see where to put a track for 33k 2w resistor it would go top side of board to right terminal, does the other one go from left side of D2 again on top side around outer edge of board past C1 and ZD1 to left terminal ?

 

[alec_t]

Something like this? Like Tina, Eagle doesn't allow vertical mounting of diodes so I cheated and re-labelled some resistors.
Are the terminal positions (scaled 17mm from centreline) and pot outlines ok? The +- terminals of the meter movement are unidentified so I've put solder pads for connecting them.

 

[eddi40]

Hi Alec thnx once again , but you still havent connected the input part of the circuit to the actual terminals,also could you post the actual eagle version of this so i can open it up in eagle to read it better,and also post the schematic as well. so i can print it out and open it up in eagle to play with it if i have to. Its now looking good. Apart from the 78L05 and thermistor and a couple of diff value caps i think expense is better, do you think this will be final version? I hope so , im not looking forward to telling my guy tomorrow, i hope he has not worked on it really , he cant get mad at me then.The thermistor will have to be a leg version im gonna cost it up later tonite. Thankyou Alec for your effort i will let you know how it goes and if there is any probs ,if there are none it will be another circuit to add to your list , i wanna have a go at making some of em sometime. Just give me a final list of components and what type especially caps (ceramic disc, bead, electrolytic,etc etc).I need to know for sure i dont wanna waste any money now im this close! Is those 39nf type i showed you in an earlier post still ok for this.

 

[eddi40]

https://uk.rs-online.com/web/p/thermistor/5288508/
hi will this thermistor be ok couldnt find a 3.3k version of a leaded ntc thermistor. Is 300r less critical ? Also you said if it had the thermistor on the board you would eliminate the regulator , i see both are on this board. Is this because the board is using both 4093 and 4013?

 

[alec_t]

you still havent connected the input part of the circuit to the actual terminals

It wasn't clear at the time whether the terminals were those of the meter movement itself (coil), or of the input to the circuit; so I left them for you to connect. I've now connected them.

will this thermistor be ok

Looks fine.
Re the 39nF caps, what size are they (I had to guess)?
I'll draw up a list of components to be used, and post the Eagle files. Fingers crossed that will be the final version (though personally I'd have gone for the one in post #42 )

 

[eddi40]

hi Alec thnx again, i cant see the meter input terminals on the board can you please put a couple of lines off board pointing to the pads labeled + and - . The caps have a lead spacing of 5mm ,thankyou again. Also how can i print out your drawing so that it is correct scale, i want to stick it to a piece of double sided board so i can drill thru the center of all pads then add lines to produce a nice neat board. I want to make a good job of the board now you have gone to all this trouble. That is the best way to get it accurate.

 

[alec_t]

Also you said if it had the thermistor on the board you would eliminate the regulator , i see both are on this board. Is this because the board is using both 4093 and 4013?

I was trying to make the board versatile to allow options; hence places for the regulator and thermistor. However, I've been doing a bit more simulating and with the 4013 used, as you wanted, a 5V supply won't quite allow the 4093 to drive enough meter current for full-scale deflection. Even without the 4013 it's marginal.
So in the attached I've eliminated the regulator and shuffled the components round a bit. The 4013 can now be used (optional), and R1 can be increased to 56k to reduce heating by a factor of two. It should now be ~ 0.7W; but still enough unfortunately to heat up the inside of the meter (40C?). In simulation the thermistor should cope with that pretty well, but the optimum value for series resistor R4 (post #42) depends somewhat on the setting of pot VR2. Hence the revised board has a pot VR3 in the position where R4 was (but you could try R4 fixed values ~8.2k if you prefer).
I've attached the Eagle files. See the Note on the sch file. You can print the board layout 1:1 from within Eagle. Some of the track spacings (particularly within the IC outlines) will be tight. You may have to re-route those tracks. I could only find SMD footprints for the pots, so you'll have to adjust those.

I make the components:
1 x CD4013
1 x CD4093
1 x BZX84C8V2L or other 8V-12V zener
3 x 1N4148 or similar
1 x 3k thermistor
1 x 56k 2W
1 x 22k
1 x 10k
1 x 100k trimmer
2 x 10k trimmer
1 x 47uF 25V electrolytic
1 x 0.1uF (= 100nF) plastic
1 x 39n plastic or ceramic

Good luck with the build.
When calibrating the meter, do so at whatever average temperature you think the inside of the meter will reach when in use by the customer if you can. I'd guess ~ 30-40C.
Adjust the pots in the order VR1, VR2, VR3. Vr2 and VR3 will interact somewhat.

 

[eddi40]

hi . i just moved some of the components a little just to avoid each component warming the next 1, im posting it here just so you can cast a glancing eye over it and make sure i havent screwed it up . I dont think i have i had original open using my mac while modding it on my pc right next to it so i could keep an eye on both,that eagle software looks good i must try to get used to it.If you think it looks ok ill go ahead and start to get other parts and build it up . Where you have red for top track and blue for bottom can i swap it around as i dont have ability to make thru hole connections for soldering but i can make vias by putting a small piece of thin wire thru board and soldering it to top and bottom copper layer. As chips are surface mount they can be soldered as SM then put a via next to each leg that needs to go under, ithink itll work what do u reckon?. To bridge the legs on the chips as they are so small i was going to just put blobs of solder across pins , i think theyll stand the heat for the short time they get hot, otherwise ill try to get legs shorted by drawing them onto the copper , i dont know yet.

 

[alec_t]

Alas, Eagle (Lite 5.6.0) won't open that file and reports 'Invalid data in file'.

Thinking about safety, R1 will have around 200V RMS (280V peak) across it so it must be suitably rated (300V or more). Also, the right hand end of R1 will be ~200V with respect to the rest of the circuit so its connecting track and body should be at least 3mm from anything else. My layout therefore should be adjusted to move the trimmers away from R1's track and move R1 away from the board lower edge. When working on the circuit with power applied take all due precautions to prevent electric shock. 200V is potentially lethal.

to avoid each component warming the next

R1 is the only component which generates any significant heat. Come to think of it, it could probably be replaced by an X2 rated capacitor to eliminate the heat source. I'll play with LTSpice, see what's possible, and get back to you.

that eagle software looks good i must try to get used to it

I'm still at the learning stage with it!

Where you have red for top track and blue for bottom can i swap it around

Yes. If you right click on anything and select 'Properties' you get a dialog box where you can specify the layer the object (track etc) is in.

As chips are surface mount they can be soldered as SM then put a via next to each leg

Good idea.

blobs of solder across pins

That will work. Bridging of pins is hard to avoid usually; for once it will be useful.

Bear in mind that the GND connection in the schematic is circuit common rather than true protective earth (ground). I don't know if the customers's 200V (or your test supply) has one terminal grounded or if it's floating. Make sure you don't get your wires crossed or there could be a bang!
The circuit needs electrical screening. Presumably the meter casing is metal and will be grounded in use?

 

[alec_t]

Good news. R1 can be replaced by a 10nF (0.01uF) 275VAC X2-rated capacitor. (Size 9 x 10 x 4 mm. Pin spacing 7.5mm for the BCcomponents brand).

Edit: ....but a 1k resistor (ordinary 0.25W low voltage type) is needed in series with it to protect the circuit from switch-on current surges.
Revised Eagle files attached.
Power consumption reduced from 700mW to 28mW. So meter interior temp won't be much above ambient and residual drift should be reduced too.

 

[eddi40]

ohhhhh Alec ,alec,alec, components were ordered today please tell me you havent removed the thermistor, that was an expensive item and its been ordered today 6 of em at £1.38 each my pocket aint very deep. Also 56k 2W resistors were ordered too, how much are these capacitors? and if the resistor is replaced by this cap and thermistor is still on board it wont matter what temp unit is calibrated at will it? The room temp averages 19.5-20 degC. The case of instrument is aluminium and the terminals are insulated from a brass back by tufnol bushes and there is a soldered earth bush on the base with a tag screwed to it so customer can earth it. The entire instrument is 500v insulation tested and 1000v flash tested both tests are performed between each terminal and earth on base, if it goes LIVE it will go straight to Natural Earth so dont worry about that and yes i do know about the voltages i have made 2000v A/C m/c instruments before using 2M high voltage resistors before but only problem with that job was the fluke5101b calibrator would only go up to 1200v so we had to set instrument a little high on the magnetizer block as 1200v was just over half scale so we could allow the top reading to be right when customer got it. I know this works cos he asked for 5 off on the first order and i had to make 1 same style as this one but not seal it in case , send it to him if he was happy with its accuracy he sent it back , i sealed that one up and built 4 more exactly the same way. To date i have done 10 now, not a big fan of guesswork calibration but sometimes thats how its got to be done if equipment doesnt go high enuff on the voltage side 2Kv voltmeter is the highest i can make. Also d/l (EAGLE 6.1.0 windows) that is the version i used to alter the tracking a little bit, it might open then. I guess as temp is not much of an issue now i can swap out Vr3 back to the 8k2 0.5w resistor which should make it a bit easier to calibrate having a third pot i thought was going to make it a bit fiddly, playing 3 pots off against each other till readings were right could have taken a very long time. what do you think can i remove Vr3 and put back the 8k2 0.5W resistor ? If its not that value can you tell me what it is , please dont bother to post another schematic you have done so well for me so far i dont want you to spend more time uploading more schematics unless you think it will help other people who may be watching for the result of this.My mate has come up with a solution for the SMD chip tracking problem for initial board design, where there is 4 pins on one side all joined together hes is just going to draw 1 big block for those, and blocks of 2 for the others that have 2 joined and on the one chip 2 legs dont even go anywhere so hes's not going to do any pads for it, then solder necessary pads to each big block. He's a genius he has built his own fishing bite alarm which fits into his fishing reel, clever ehh. I told him to make a few and sell em on ebay but he wont. Even patent it and give another firm the headache of sales and marketing and just cream in a percentage but he wont. Anyways let me know on the 8k2 0.5W res thingy. WILL THIS ONE WORK:- **broken link removed** Let me know ASAP so i can buy them also those ceramic 39nf i showed you are they still suitable.

 

[alec_t]

components were ordered today

Ah. An unfortunate consequence of starting this design after the project was overdue for delivery .

please tell me you havent removed the thermistor

It's still there .

how much are these capacitors?

Farnell have them listed at £0.30 each.

thermistor is still on board it wont matter what temp unit is calibrated at will it?

Not really, unless you want to optimise the temp compensation.

do you think can i remove Vr3 and put back the 8k2 0.5W resistor ?

Ok, but be prepared to swp 8k2 for a slightly different value. That's the value the simulation gives, assuming the thermistor is exactly 3k and has a standard temperature coefficient (beta) and the zener voltage is exactly 8.2V. A trimmer would have allowed better for component tolerances.

WILL THIS ONE WORK:-

Don't know. The voltage rating isn't shown; and that's critical. It must be a good bit above the 200V RMS (I've assumed all along that 200V means 200V RMS ?) 280V peak.

those ceramic 39nf i showed you are they still suitable.

Yes.

 

[eddi40]

it is shown, they are 275v the same as the others you said, i was going to use the 10v zeners you mentioned in a previous post. Can i calibrate this instrument using Vr1 and Vr2 and just use Vr3 at the end just to fine tune it.

 

[alec_t]

it is shown, they are 275v

Sorry. Missed that.

i was going to use the 10v zeners

Fine.

Can i calibrate this instrument using Vr1 and Vr2 and just use Vr3 at the end just to fine tune it.

Yes. Initially set Vr3 at 8k2 then adjust Vr1 for 0% reading at 400Hz. Then set Vr2 for 100% at 440Hz. If Vr3 is too high the meter reading will increase with temperature (under-compensation); if too low then the reading will decrease as temperature rises (over-compensation).

 

[eddi40]

hi dont you mean adjust Vr1 for 0% at 360Hz,or 360hz on the existing dial, and movement has to be set at 100uA exactly at FS on Dial with a coil resistance of exactly 2k R, before it is connected to the circuit board.Also the other thing is do i have to set readings at 360 then 440 then put a case on and leave it running for a few mins,otherwise how will i know if reading is going to rise or fall and what happens if i leave it for 5mins to warm up then go back to it retune it using vr3 and then leave it another 5 mins and it rises or falls again i could be there all day trying to get it right. What happens then, cos as far as i will be concerned it will be just as unstable as previous version. This temp thing seems a bit hit and miss. Let me know your thoughts.

 

[alec_t]

dont you mean adjust Vr1 for 0% at 360Hz,or 360hz on the existing dial

Yes. Brain not in gear again. I assume movement is 2k, FSD is 100uA, and the scale reads 360Hz with 0uA (=0%), 440Hz with 100uA (=100%). If the movement isn't exactly 2k/100uA then the Vr2 adjustment should compensate for that.

i have to set readings at 360 then 440 then put a case on and leave it running for a few mins

Yes.

i could be there all day trying to get it right

I don't think so. One or two twiddles of Vr3, at 5 min intervals as you suggest, hopefully will suffice.

as far as i will be concerned it will be just as unstable as previous version

Well simulation shows it should be more stable. But simulation is just that. The real world can be different. It's a chance that has to be taken as I'm not in a position to actually build this circuit so can't offer any guarantee on its workability. All I can say is that other circuits designed using LTSpice have performed in practice very closely to their simulation.

 

[eddi40]

ok thanks Alec double sided circuit board has been ordered i am now getting close to getting this built i will let you know the results. If its left unchecked for ,lets say 30 mins when i reconnect it will i have to wait a couple of mins for reading to rise or fall to its correct position on dial or will it read correctly within a few seconds and stay on the reading? Or is that something i will not know until its built and calibrated ? Will let you know , thank you for all your help . I will keep you informed.

 

[alec_t]

If its left unchecked for ,lets say 30 mins when i reconnect it will i have to wait a couple of mins for reading to rise or fall to its correct position on dial or will it read correctly within a few seconds and stay on the reading?

When you power it up it should give a steady reading within a fraction of a second. Until you've got Vr3 adjusted correctly the reading will then drift slowly (hopefully only slightly ) with any temperature change (which shouldn't be much if the X2 cap is used, R1 is only 1k and ambient temp change is small).