DCLoad - My first analog project. Some Questions, help?

[cthree]

I designed my first pure analog electronic project and I would appreciate a general review by experienced eyes and I have a couple of questions I'm hoping to answer as well.

Disclaimer: I didn't "design" this so much as mashup some designs already out there so I don't claim any sort of credit. When I'm done and I've build and tested it I will release it to the public without significant restriction via github and make the PCB available via OSHPark etc.

My first question relates to Q4 and Q5. I don't have any P-channel mosfets to breadboard this up but what I'm doing there is using the latching input switch to toggle between displaying the load set point when the input switch is low/gnd (via Q4) and displaying the actual load when the input switch is high/5Vish (via Q5). Does this look valid and sensible to experienced eyes?

Second, is the Modulator Input section which I'm a little stumped on. I want to drive the load via a function generator signal which I can sort of do by AC coupling the signal with a cap but that's not sufficient. What I want to achieve with the modulator input is to have the main pot set the DC bias and then modulate the AC signal from the function generator on it. Where I'm stuck is scaling the AC signal so it retains it's input waveform but to attenuate it so that is stays within the safe operating range regardless of the input signal. I'd rather not feed in a 30Vpp signal and blow up my opamp if you know what I mean. Can you point me in a direction to learn about how to implement this feature? The design inspiration for the idea used a simple resistor voltage divider which I don't like for a number of reasons including not handling the negative phases and causing a 90 degree phase shift. Ideally I want the signal phase to remain true to the source or things could get confusing quickly.

I've calculated that I can dissipate 25W through the pass transitor Q1 using the heatsink and fan I have specced. I've chosen 5A and 30V as absolute max. owing to the capacity of the 30VDC/5A relay K1 which means U2-2 shouldn't see more than 500mV on either input (100mV/A). The modulator would modulate the  input from the buffer U2-1 and main 10T pot VR2

Third, can I run traces between two pins (K1 and Q1) on both sides of the PCB to double the current capacity? On the PCB I'm using a 50mil trace on the bottom side (without solder mask over the trace) and I'm wondering if I can run a trace on the top as well. I'm not sure what spooky effect this might have.

That's it for now. There are a few known issues like R13 and R14 being the wrong values. I've deferred that until my panel meter gets here on the slow boat from China. It's a CX102B from Futerlec. The opamps U2 and U3 are TLC272ACP dual opamps I have in quantity on hand. Your feedback and comment most appreciated.

Thanks!

[kizzap]

the first thing I can see... you need to properly terminate U3-1. Generally done by bringing a voltage half v+ to the non-inverting input, then having the output feed into the inverting input. It stops that op-amp playing around. (I think you might be wanting to use it in the waveform input stuff?)

I'm sceptical about the need for the on-board regulator. DC plug-packs are so cheap these days, and it even looks like you should be able to get the design to work off one rail.

My first question relates to Q4 and Q5. I don't have any P-channel mosfets to breadboard this up but what I'm doing there is using the latching input switch to toggle between displaying the load set point when the input switch is low/gnd (via Q4) and displaying the actual load when the input switch is high/5Vish (via Q5). Does this look valid and sensible to experienced eyes?

To me, it sounds like you could do what you are doing with the 555 timer with a simple SPST button. The output of the 555 is just a toggle output, so you could simply put the button in there so that when the button is open, it shows one state and when closed, shows the other. This does mean that you would have to hold the button to consistantly see one display, but if that is the maximum set, I don't see that being an issue.

Second, is the Modulator Input section which I'm a little stumped on. I want to drive the load via a function generator signal which I can sort of do by AC coupling the signal with a cap but that's not sufficient. What I want to achieve with the modulator input is to have the main pot set the DC bias and then modulate the AC signal from the function generator on it. Where I'm stuck is scaling the AC signal so it retains it's input waveform but to attenuate it so that is stays within the safe operating range regardless of the input signal. I'd rather not feed in a 30Vpp signal and blow up my opamp if you know what I mean. Can you point me in a direction to learn about how to implement this feature? The design inspiration for the idea used a simple resistor voltage divider which I don't like for a number of reasons including not handling the negative phases and causing a 90 degree phase shift. Ideally I want the signal phase to remain true to the source or things could get confusing quickly.

The two things I can think of are either a biased transistor amplifier, or a summing amplifier.

I've calculated that I can dissipate 25W through the pass transitor Q1 using the heatsink and fan I have specced. I've chosen 5A and 30V as absolute max. owing to the capacity of the 30VDC/5A relay K1 which means U2-2 shouldn't see more than 500mV on either input (100mV/A). The modulator would modulate the  input from the buffer U2-1 and main 10T pot VR2

Be careful here. I would be cautious on the specs of the relay to begin with, however if you have a limit of 25W, then you need to be able to limit both voltage and current flowing through the pass Transistor. If you don't, you could suddenly be pushing 150W + through your pass element.

Third, can I run traces between two pins (K1 and Q1) on both sides of the PCB to double the current capacity? On the PCB I'm using a 50mil trace on the bottom side (without solder mask over the trace) and I'm wondering if I can run a trace on the top as well. I'm not sure what spooky effect this might have.

Simple answer? Yes.

Complicated answer? Yes, but you need to consider the return path of the current.

That's it for now. There are a few known issues like R13 and R14 being the wrong values. I've deferred that until my panel meter gets here on the slow boat from China. It's a CX102B from Futerlec. The opamps U2 and U3 are TLC272ACP dual opamps I have in quantity on hand. Your feedback and comment most appreciated.

Thanks!

I am not too sure about the fan controller solution you have implemented. I'm working on a DC load myself, am I am going to be using an i5 stock heatsink (good for 90W+) and my fan control is simply reading the temperature, and switching a fan on if it goes above a set temperature level, and off when it goes under a set level. I am not sure, but it looks like your circuit might limit the speed of the fan, depending on the temperature Nope. Just looked at it further. It is literally going to do nothing, purely due to R7. If R7 is not there, then Q2 might be able to control the fan.

-kizzap

[cthree]

Many thanks for your reply.

Some of my choices are biased towards parts on-hand. The use of the 555 for instance. I have a whole tube of 555s and about 5 RA momentary buttons but no pushbutton toggle switches so I agree with you there. U3-1 is a work in progress, I'm reserving it for the modulator function.

I'd love to add protection and limits on the input to clamp it at below 25W. All I want is to sink 3A@5V and 5A @3.3V really. As soon as I figure out how to limit it to 0-5A and 0-30V and 0-20/25W I will add that circuit in there. For now I was just going to be aware of the limits because I wouldn't be near them.

My bad. R7 is a load option to omit all the fan control stuff and just make the fan go one speed. The HI trimpot sets the turn-on hysteresis and the LO trimpot sets the turn-off hysteresis. This part I have prototyped and it's working great.

I need the 5V rail for the panel meter and the voltage reference. The TLC272 can only work to Vcc-1.5V and 3.5V isn't enough to fully turn on the IRF530N MOSFET. If I were to use a IRL530 MOSFET which has a much lower Vg(th) you are probably correct about using just a 5V rail. I also have a full tube of LM317's so it seemed sensible and convenient to use 12V and 5V rails. The voltage reference is there to keep the opamp input stable and free of loading effects on the 5V rail from the display, relay and so forth. I'll go back and reconsider Q1 and all that.

Thanks again!