Maker Pro
Maker Pro

What value of resistors do I need and a sanity check

KrisBlueNZ

Sadly passed away in 2015
RL and RB are used to provide currents of around 1 mA for the switched LED and the voltage reference LED. The values have to be adjusted according to the battery voltage used. These changes could be avoided if they were replaced with constant current sources.

There used to be the 1N5283 series of "current regulator diodes", which were really just JFETs with the gate and source tied together, specified to pass a particular maximum amount of current. These seem to be discontinued, or only available in huge quantities.

Does anyone know of a similar simple current regulator device (or a JFET designed to be used for this function) that is available in small quantities with a 1 mA current specification?

Edit: It looks like a simple answer would be a JFET with its gate grounded, and a resistor from source to ground. The JFET will self-bias so the source is at a (roughly) fixed voltage above ground, and you adjust the source resistor to get the desired drain-source current. The NXP J111 or J112 looks suitable. Comments?
 
Last edited:
Sorry for the relentless questions, but Kris could you post a link to the specific trimpots and resistors. I went through the kits already, but i felt like some of it was in Japanese it was difficult to understand; as long as the price doesnt exceed around 60-75$ it would be great. Also if its cheaper could provide the link to the 10 turn trimpots.
 

KrisBlueNZ

Sadly passed away in 2015
It would help if I knew which country you were in. I'll assume you're in America and point you to the Digikey web site. Digikey and Mouser will sell you almost any electronic component you would ever want.

Here's an updated schematic.

attachment.php


Resistors in this application are not critical. For normal applications I use 1% metal film parts. These are made in the "axial" (cylindrical with a wire out each end) style. Go to http://www.digikey.com/product-search/en/resistors/through-hole-resistors

Digikey have three suitable ranges that are available in low quantities:

Vishay "SFR16S" series, 0.5W. One-off quantity but most expensive.
Vishay "SFR25" series, 0.4W. One-off quantity. Probably the best option.
Yageo "MFR" series, 0.25W. Minimum order quantity is 5.

Narrow down your selection using the filter table. Select the series you want, check "In stock" and click "Apply Filters". If you chose the SFR25 series, you will now have the option of 100ppm/C or 250ppm/C temperature coefficient. Most of these resistors are 100ppm/C so select that and reapply the filters.

When you're using the Digikey filter table, you can use Shift-click to select a range, and Ctrl-Click to select/deselect individual lines in a multi-line selection. Also, ignore any lines where the minimum order quantity is more than 5, or that say "Non-stock". Yes I know you checked "In stock" but they still sometimes list those lines. I don't know why.

You can now see the resistance values available in the SFR25 series. I'll come back to the Digikey page later after the nominal value calculations.


While the circuit is in regulation, i.e. providing the selected current to the load, there will be a roughly constant voltage across the emitter resistance. That is, a roughly constant voltage between the positive supply rail and Q1's emitter. The output current can be calculated using Ohm's law, which says I = V / R. In this case, I is the output current (in amps), V is the voltage across the total emitter resistance (in volts), and R is the total emitter resistance (in ohms).

V is equal to the forward voltage of LEDC (when operated at a forward current of about 1 mA) minus the base-emitter voltage of Q1 when operated at a collector current of 0.5~2.5 mA. The Avago data sheet says the LED has a nominal forward voltage of 1.6V at 1 mA, but this is only nominal and there is variation between individual LEDs. I'm assuming 0.6V base-emitter voltage for the transistor, which is reasonable considering the low base and collector currents that it will be operating at. The actual voltage will depend on the transistor type, and characteristics of the individual transistor. These parameters will also vary with temperature; you should test the unit at low and high temperatures to see how much the current varies.

With those values, there will be 1.0V across the emitter resistance. I have marked this on the updated schematic.

This means V = 1.0 in the Ohm's law formula. Therefore the output current (in amps) is 1.0 divided by the total emitter resistance (in ohms). Ohm's law can also be rearranged to R = V / I which means that R (ohms) = 1.0 / I (amps).

First we have to calculate RX and RS. These determine the maximum output current. They are in series, so their resistances add together, and we calculate assuming the trimpot will be at mid-position, so its actual resistance will be about half of its total resistance. For example a 100 ohm trimpot set half-way will have a resistance of about 50 ohms. (The trimpot is there so you can set the current exactly at room temperature.)

You want the maximum output current to be 2.5 mA, which corresponds to a total emitter resistance of 400 ohms (from R = 1.0 / I). So RX+(RS/2) should be 400 ohms. Most (say 80-90%) of the resistance should be in the fixed resistor RX, with the rest in the trimpot RS.

Trimpots are generally available in the 1-2-5 series, e.g. 10, 20, 50, 100, 200, 500 etc. Expensive trimpots are available with higher accuracy and value resolution but I'll stick with these options.

So RX + (RS/2) = 400 ohms, and RX should be a large percentage of that amount.

Looking through the available resistances on the Digikey page, I see 348 ohms is available. This is 87% of the total resistance needed, and it fits nicely with a 100 ohm trimpot.

RX = 348 ohms (Vishay SFR25 series)
RS = 100 ohms (multi-turn trimpot)

Now the four preset currents.

I've reconsidered my suggestion to use a resistor and a trimpot in series for each of the four preset currents. A single multi-turn trimpot for each will be simpler. I've updated the schematic to show this. I would like to stick with the fixed resistor RX from the emitter for safety reasons - it limits the maximum current in case of a short somewhere in the emitter resistor circuit (you must protect the connection from RX to the emitter of Q1 so NOTHING can short onto it).

Each of the four fixed currents will be set by one multi-turn trimpot (selected by the switch) in series with the 400 ohm resistance made up from RX and RS, so we have to subtract 400 from the value that we calculate based on the output current. Also, we want each trimpot to be set around 2/3 to 3/4 of its maximum resistance, so its rated resistance needs to be between 130% and 150% of the value we calculate.

2.0 mA: R = 500; RU at 2/3 positon = 100; RU = 200 ohms (multi-turn trimpot)
1.5 mA: R = 666; RT at 2/3 position = 266; RT = 500 ohms (multi-turn trimpot)
1.0 mA: R = 1000; RS at 2/3 position = 600; RS = 1000 ohms (multi-turn trimpot)
0.5 mA: R = 2000; RR at 2/3 position = 1600; RR = 2000 ohms (multi-turn trimpot).

Trimpots are at http://www.digikey.com/product-search/en/potentiometers-variable-resistors/trimmers

There are two series that look suitable. Both types are 12-turn, 1/4 watt 10% cermet types and are available with top or side adjustment.

Murata "PV37" (wide variety available)
Bourns 3266 (more expensive).

Go to that web page and select the series you want, check In Stock, and click Apply Filters. Now you can choose the values you're interested in - 100, 200, 500, 1000 and 2000 ohms, and apply filters again.


Finally you need a value for RV. When it's fully clockwise, the current will be 2.5 mA as set by RX and RS. When it's fully anticlockwise, the current will be set by the resistance of RV plus 400 ohms from RX and RS.

Say you want the minimum variable current to be 0.3 mA. The total emitter resistance needs to be 3333 ohms (from R = 1.0 / I). Subtract 400 ohms for RX+RS and you get 2933 ohms. A 3k potentiometer would do, except that potentiometers usually have 20% tolerance on their end-to-end resistance, so you might not be able to get the current all the way down to 0.3 mA.

As you can see, you won't be able to adjust the current right down to zero.

How about choosing a 2k potentiometer. This gives a current adjustable down to 0.41667 mA nominally. If that's low enough for you, go for it. If you make RV much higher, you can get the current lower, but you'll find that the most useful part of the adjustment range, from say 0.5 to 2.5 mA, is over a very small angle of adjustment. So there is a compromise there. If you really need a wide range, you could add another position on the rotary switch and use two potentiometers.

As for sourcing the potentiometer, I'll leave that to you, OK? :)

Edit: You should also get an adjustment tool for the trimpots. You can adjust them easily enough with a flat blade screwdriver, but the buggers keep sliding out of the slot and falling off! There is a proper tool for this, known as a swizzle stick tool, trimpot screw adjuster tool, or trimmer adjustment tool. See
http://www.digikey.com/product-detail/en/ACCTRITOB308-T000/SPTOOL-ND/1287467
 

Attachments

  • Jjanes.gif
    Jjanes.gif
    13.6 KB · Views: 356
Last edited:
I understand the website messes up sometimes but I need an alternative to the diode DZ because the website will not let me buy a smaller quantity than its bulk package.

what would you recommend?

thanks
 

(*steve*)

¡sǝpodᴉʇuɐ ǝɥʇ ɹɐǝɥd
Moderator
Curious. DZ could be there for a number of reasons, but I'm pretty sure that its reverse voltage isn't one of them...

KrisBlueNZ may have to explain exactly why it's there.

On that basis I won't offer a suggested alternative.
 

(*steve*)

¡sǝpodᴉʇuɐ ǝɥʇ ɹɐǝɥd
Moderator
AH!

DZ is connected across LEDC to limit Q1's base bias if LEDC becomes broken or disconnected, to prevent excessive output current. DR protects Q1 against connection to some types of inappropriate loads. F1 is a tiny 5 mA fuse intended as a simple safety precaution in case the current regulator fails in some way (for example, Q1 is damaged).

OK, it protects against the LED going open circuit and also (possibly inadvertently) protects the transistor (but not DR or DZ) if the batteries are connected the wrong way around.

I guess Kris is most concerned that the LED, if mounted away from the PCB, might become disconnected.

You could mount a higher voltage LED (say a green one) in parallel with the red one, but mounted on the board so it is safer. If the red LED were disconnected, the greed LED has only a slightly higher Vf and would take the place of the zener for that particular fault.

See what Kris thinks...

Where are you getting your components from anyway? Perhaps there is another alternative that doesn't have a minimum order quantity?
 

KrisBlueNZ

Sadly passed away in 2015
OK, it protects against the LED going open circuit and also (possibly inadvertently) protects the transistor (but not DR or DZ) if the batteries are connected the wrong way around. I guess Kris is most concerned that the LED, if mounted away from the PCB, might become disconnected.
Good deduction Sherlock :) It's a product of my creeping paranoia about the whole idea of this application. The reason is buried in the description in post 52.
You could mount a higher voltage LED (say a green one) in parallel with the red one, but mounted on the board so it is safer. If the red LED were disconnected, the greed LED has only a slightly higher Vf and would take the place of the zener for that particular fault.
That's probably just as valid. The most important thing is a sharp knee and low leakage voltage, because when the LED is there doing its thing, you don't want the zener drawing any significant current, but then you want it to take over without letting the voltage across it rise too much, because that rise corresponds to proportionally higher output current. Steve, what would you recommend here?

I had another idea to use some of the current from RB to bias up a few 1N914s in series and use them to provide a voltage reference, with a diode from Q1 base to the reference voltage to limit the voltage. That would give a very low leakage and a sharp knee. Perhaps I'm being too paranoid?
 
I found a diode similar to DZ. Its difference is that instead of 1.5V @ 100mA it is 1.5V @ 200mA. It also has 5 tolerance (three fewer than DZ) and 70 fewer Ohms in imped. resistance.

What do you all think?
 

CDRIVE

Hauling 10' pipe on a Trek Shift3
Good deduction Sherlock :) It's a product of my creeping paranoia about the whole idea of this application. The reason is buried in the description in post 52.

Which is why I bailed from this topic early. This thing is a liability attorney's wet dream. :eek: Actually, this is only partially true. If there isn't a deep well to suck dry they're not interested at all. Anyone here have a deep well? :rolleyes:

Old Canadian expression: CYA! ;)

Chris
 

KrisBlueNZ

Sadly passed away in 2015
Which is why I bailed from this topic early. This thing is a liability attorney's wet dream. :eek: Actually, this is only partially true. If there isn't a deep well to suck dry they're not interested at all. Anyone here have a deep well? :rolleyes:
My financial situation is reminiscent of a dried-up watering hole on the plains of the Sahara desert LOL
i meant to mention that that difference is its forward voltage. its leakage value stayed the same 100µA @ 1V
That's too much leakage current. It needs to be 100 uA or less at 1.6V (that's 1/10th the current of the LED) otherwise it will affect the circuit's normal operation too much. I could change the circuit to use several diodes with a separate bias path but it seems like overkill and it might be simplest to omit DZ and construct the circuit in such a way that it's very hard for the LED to become disconnected. There is a 5 mA fuse on the output, which will protect against excessive overcurrent.
 
I could change the circuit to use several diodes with a separate bias path but it seems like overkill and it might be simplest to omit DZ and construct the circuit in such a way that it's very hard for the LED to become disconnected. There is a 5 mA fuse on the output, which will protect against excessive overcurrent.

If it makes it any easier, i could care less about it having the led design that jjanes tried to attach. In fact if its easier i would preffer for the device to just be one that simply delivers the 2mA. If you want to add as many safety precautions as you want, go ahead.

If anything bad were to happen I Fernando cannot and will not hold anybody accountable. All i really want is help with this.
 

KrisBlueNZ

Sadly passed away in 2015
Using the LED as the reference voltage is a very good idea because the LED also indicates whether the current is flowing. If the desired amount of current is not flowing - for example if the electrodes aren't properly connected, the LED goes dim or goes out.

You can omit the zener if you want. It's only there for safety. Just make sure the connections to the LED can't come disconnected. For example, don't solder it straight through the stripboard (I assume you'll be using stripboard) because if something gets bumped, it's possible for the copper strip to be ripped off the stripboard and the LED would become disconnected.
If you only want a fixed 2 mA current, you can eliminate the rotary switch, connect RM to the positive supply from SW1, and use RX around 400 ohms (390 ohms should be available in that resistor series) and RM = 200 ohm multi-turn trimpot.
 

KrisBlueNZ

Sadly passed away in 2015
I've been asked to write up a parts list. This list matches schematic revision 002 (apart from the RV value, see below).

BT1 Battery 12VDC
-- whatever you want to use. A battery holder with AAA or AA cells is probably a good idea.

DZ omit


DR 1N4007
http://www.digikey.com/product-detail/en/1N4007/1N4007FSCT-ND/965481

F1 0279.005V 5mA Littelfuse
http://www.digikey.com/product-detail/en/0279.005V/0279.005V-ND/553339

LEDC HLMP-D150 (5mm)
http://www.digikey.com/product-detail/en/HLMP-D150/516-1323-ND/637587

LEDR,S,T,U,V 5x HLMP-K150 (3mm) x5
http://www.digikey.com/product-detail/en/HLMP-K150/516-1311-ND/637575

Q1 BC557B, 2N3906, MPSA55/56, 2N5087, 2SA1309
http://www.digikey.com/product-detail/en/BC557BTA/BC557BTACT-ND/3042484

RB 10K Vishay SFR25
http://www.digikey.com/product-detail/en/SFR2500001002FR500/PPC10.0KYCT-ND/596847

RL 8K2 Vishay SFR25
http://www.digikey.com/product-detail/en/SFR2500008251FR500/PPC8.25KYCT-ND/597164 (this one's actually 8k25; difference is unimportant)

RM trimpot 100R
http://www.digikey.com/product-detail/en/3266W-1-101LF/3266W-101LF-ND/1087900

RR trimpot 2K
http://www.digikey.com/product-detail/en/PV37W202C01B00/490-2975-ND/666602

RS trimpot 1k
http://www.digikey.com/product-detail/en/PV37W102C01B00/490-2969-ND/666596

RT trimpot 500R
http://www.digikey.com/product-detail/en/PV37W501C01B00/490-2982-ND/666609

RU trimpot 200R
http://www.digikey.com/product-detail/en/PV37W201C01B00/490-2974-ND/666601

RV pot 2.5k (incorrectly marked 500R on schematic)
http://www.digikey.com/product-detail/en/91A1A-B28-B12L/91A1A-B28-B12L-ND/2564645

RX 348R Vishay SFR25
http://www.digikey.com/product-detail/en/SFR2500003480FR500/PPC348YCT-ND/597022

SW1 SPST slide or toggle
-- slide: http://www.digikey.com/product-detail/en/S101031SS03Q/CKN10054-ND/2043289
-- toggle: http://www.digikey.com/product-detail/en/7105D/451-1158-ND/2679541

SW2 2-pole 5-position rotary
http://www.digikey.com/product-detail/en/R20507RR02Q/CKN9483-ND/1949304

M1 3mA DC optional
-- This meter shows the current flowing in the output circuit. It's desirable but not necessary. I suggest a mechanical moving pointer type meter for this position. I can't find anything suitable at Digikey or Mouser.

I recommend you get a trimpot adjuster
http://www.digikey.com/product-detail/en/ACCTRITOB308-T000/SPTOOL-ND/1287467

You'll need some kind of board - I suggest stripboard (aka Veroboard in England).

You'll need output connectors - for example, "banana" plugs and sockets, and a case (use PLASTIC).
 
Thank you Chris for the parts list. Very helpful. Thank you to everyone else that has contributed to this project. Even though this is a simple circuit, I wanted to make a sim for it. What I used was MultiSim. It seems to work good as far as I can see besides the LEDC which I had to low the mA to .95 in order to turn on. I don't know if this is correct or why this would be when that LED should be at 1 mA. Could anyone look this over or run this to make sure I did everything correct. Thank you.

Here is a link to my dropbox of the .ms11 file.

http://dl.dropbox.com/u/15767300/tdcs finish.ms11

Again thank you.

Best,

Bobby
 

KrisBlueNZ

Sadly passed away in 2015
I don't know what to do with an ms11 file. Maybe you could upload a screen shot of the simulation and results. I should be able to check that.
 
Hi.... Sorry.. Kris... didn't mean to spell your name wrong... Here is a screen shot of the circuit. If there is any other information you need to complete this please let me know. I tried as close as I could to stay with the same layout as your circuit is on. Thanks...

Best,

Bobby
 

Attachments

  • tDCs.jpg
    tDCs.jpg
    148.3 KB · Views: 576

KrisBlueNZ

Sadly passed away in 2015
That looks good, Bobby.

Regarding DZ. Can you tell from the simulation, how much current is flowing in DZ? If it's less than around 100 uA it should be OK to leave it there, but if it passes too much current, it will affect the base voltage, which is supposed to be determined by the LED only. (The zener was intended to be a safety feature in case the LED became disconnected.)

If DZ is drawing more than 100 uA, I think it's simplest to omit it. Make sure your connections to the LED won't break if bumped or pulled. If the LED becomes disconnected, the current generator output current could increase greatly. The output fuse will blow if the output current significantly exceeds 5 mA for long enough.
 
Last edited:
Top