Warning LED when valve stops conducting - a little SS help needed

[RhythMick]

Thinking about this some more, I don't like the first solution at all now, I have gone off it, I would prefer something like the second, which isn't in the music signal path. Try and LED in series with 820R between the base of the D44H11 and ground and see if the LED lights up and then check if the valve heater makes full voltage. Measure what full is for the valve before and after doing the change.

Then have a listento make sure I haven't broken anything.

Andrew

As always very much appreciate the time and input fellas. I feel some experimentation coming on, maybe even tonight.

[RhythMick]

Thinking about this some more, I don't like the first solution at all now, I have gone off it, I would prefer something like the second, which isn't in the music signal path. Try and LED in series with 820R between the base of the D44H11 and ground and see if the LED lights up and then check if the valve heater makes full voltage. Measure what full is for the valve before and after doing the change.

Then have a listen to make sure I haven't broken anything.

Andrew

Andrew let's assume I try the 820R + LED from D44H11 Base (pin 1) and ground, and it lights when the heater is conducting. Could we then change it around so the LED only lights when the heater is NOT conducting ? Instead of connecting to 820R then ground, connect to a resistor (lets say 1K for now) then to the base of the NPN in my original circuit, with a resistor and LED on the high side of that NPN.

[Andrew]

Thinking about this some more, I don't like the first solution at all now, I have gone off it, I would prefer something like the second, which isn't in the music signal path. Try and LED in series with 820R between the base of the D44H11 and ground and see if the LED lights up and then check if the valve heater makes full voltage. Measure what full is for the valve before and after doing the change.

Then have a listen to make sure I haven't broken anything.

Andrew

Andrew let's assume I try the 820R + LED from D44H11 Base (pin 1) and ground, and it lights when the heater is conducting. Could we then change it around so the LED only lights when the heater is NOT conducting ? Instead of connecting to 820R then ground, connect to a resistor (lets say 1K for now) then to the base of the NPN in my original circuit, with a resistor and LED on the high side of that NPN.

Assuming I have understood you, I can't see any situation where you could reverse the logic. In all cases, an LED would glow brighter when not conducting, but still glow when the heater was conducting. When the heater is conducting the base is at about 2v less than the supply rail and is close to ground when the filaments isn't conducting.

You'd have to drive an intermediate circuit of some kind to get the behaviour you desire, something like a transistor to drive the LED so that when the base of the transistor was pulled low by the op amp's output going low (which is also trying to swtich on the current driver for the DHT) the LED would then light. A second small NPN should do this.

EDIT: search low side switch, or similar.

[RhythMick]

You'd have to drive an intermediate circuit of some kind to get the behaviour you desire, something like a transistor to drives the LED so that when the base of the transistor was pulled low by the op amp's output going low (which is also trying to swtich on the current driver for the DHT) the LED would then light. A second small NPN should do this.

Yep - that's exactly what I had in mind. Something like this ... as I say I'm VERY new to SS so wouldn't be at all surprised if I've got something wrong here.

Hmmm ... I think I see problems with this.

Wouldn't the BC547 emitter need to be referenced to the DC board "ground" in order that the voltage at the D44H11 base can switch it ? If so does that imply taking power from the DC board rather than an external supply ?

I need to measure the D44 base voltage in both conditions ...

Valve heater conducting (expecting 2v below rail)
Valve heater not conducting (expecting near ground)

[RhythMick]

Would a comparator be a better approach. comparing the D44H11 base to it's emitter ? The comparator (and the LED) are then driven from an entirely separate supply and thus interfere less with the DC board operation. I happen to have some LM293P dual-comparator chips around, so could monitor both valves in the pair from this chip. http://www.ti.com/lit/ds/symlink/lm293.pdf

EDIT - I notice that the max output current from the LM293 is 20mA, so I need to adjust the resistor upwards to limit the current, probably to 10mA, or get the output to trigger the BC547.

When heater is conducting, D44H11 base will be 2v below rail ie above ground.
When heater is open circuit, D44H11 base will be near ground.

The above isn't quite right I need to make it react to a difference above a threshold - linking the output back to the IN- (schmitt trigger style ?) might do it.

[RhythMick]

Last attempt before tea - this is fun ...

[Andrew]

The LED shouldn't take any current when the heater is conducting, it will take a 10mA or so, when it isn't; the board's supply can deliver an amp or more...I'm not certain I see the issue, am I missing something?

Equally, why couldn't the LED and driver use an external supply? Yes, as far as I can see, admittedly after a long day, the two circuits need to share a common ground; Kirchoff and all that, but you could use a separate supply, if you wanted....

Again, unless I'm missing something, I think a comparator is overkill...you just need a reverse polarity switch.

[RhythMick]

The LED shouldn't take any current when the heater is conducting, it will take a 10mA or so, when it isn't; the board's supply can deliver an amp or more...I'm not certain I see the issue, am I missing something?

Equally, why couldn't the LED and driver use an external supply? Yes, as far as I can see, admittedly after a long day, the two circuits need to share a common ground; Kirchoff and all that, but you could use a separate supply, if you wanted....

Again, unless I'm missing something, I think a comparator is overkill...you just need a reverse polarity switch.

Very likely Andrew, I'm a complete SS newbie. If I can source 10mA from the board without impacting anything then the BC547 seems the way to go. Where on the board would you draw that from?

[Andrew]

The collector of the D44H11 is a clean +-ve, or on the output pin of the LD1085P, should be the same thing, if I recall correctly.

Andrew

[RhythMick]

The collector of the D44H11 is a clean +-ve, or on the output pin of the LD1085P, should be the same thing, if I recall correctly.

Andrew

Many thanks. Clip leads and meter out, will report back.

[RhythMick]

71A DC board, set for 5v output. 20R load for 0.25A. 7v RMS input.

Under load... measured to ground (output -ve)

Output +ve 5v

D44h11
1 5.7v
2 8.0v
3 5.0v

LD1085P
1 6.75v
2 8.0v
3 9.5v

Without load ... All measurements as before except...

D44H11 pin 1 dropped from 5v to 0v over approx 1 minute. Touching pin 3 with the meter caused pin 1 to rise a little, then measuring pin 3 allowed it to drain to 0v again.

LD1085P pin 3 went to 11.2v

So I think this could be a winner.

D44H11 pin 2 gives me an 8v supply. I'll use a 300R resistor to the collector of a BC547, from there to the led anode. When the 547 is off the led will light and the resistor will limit to about 19mA. When the 547 is on it should allow about 24mA.

D44H11 pin 1 via a resistor to the base of the BC547.

Emitter of BC547 to ground along with led cathode.

How to calculate the value of the BC547 base resistor? Online calculators tell me about 500R with an 8v Vcc, 20mA load and switching voltage about 1v. I don't like using those calculators because I don't yet understand whether they are doing the right thing for this application but I'll work through it later.

The 26 heater voltage is 1.5v so 1v should be ok as the switching point? Hmmm not sure I understand what that term means.

[RhythMick]

Trying to get my head around the base resistor calculation.

When the 547 is on I expect the CE current to be around 22mA. Calculated as (8v - 2x0.7v) / 300R is 22mA. The 547 should be fully saturated and the current should be limited by the resistor.

547 current gain is anywhere between 110 and 800. Work with 100 because more will just push it more into saturation.

So we want around 0.22mA on the base and I want it to switch on from around 1v at D44H11 pin 1.

(1.0 - 0.7) / .00022 = 1k4 ish. If I use a 1k I should get 0.3mA with 1v on d44h11 pin 1, which should be plenty.

Am I doing this right?

[RhythMick]

This is from the bc547 datasheet which seems to confirm.

Base current of around 0.3mA seems about right.

[Mike H]

No need to run the LED at 20 mA, 10 mA will suffice, in fact if it's a high brightness type even 5 mA is enough. So I prefer the simplicity of the first diagram without the BC547.

Secondly, why do we need the D44H11? Why not a small value resistor? A further complication not yet mentioned is how temperature changes will affect the Voltage drop of the base junction. This is a notorious problem when using diode junctions as sensors. You will have SS components in relative close proximity to hot valves. Been there and torn me hair out trying to keep things stable. A small value resistor will be much less trouble I'm thinking.

Just my 4 penn'orth.

[Mike H]

PS: when the sense resistor current drops to zero (assuming we replace the power transistor with that), you will probably need a wee bit of bias on one of the comparator input pins to make sure it switches the opposite way. this could simply be a high value resistor from the +Ve rail, for example.

HTH