Best way to make an LED turn on only when a certain voltage is reached?

Thread Starter

SiegeX1

Joined Mar 10, 2009
25
Hi Ron,

Thanks for the suggestion of using an NPN instead. My board is already using an IC that has Dual BC848 NPN's so adding one more of these IC's gives me one less part# to worry about.

Since I'll have an extra NPN, I decided to use it to drive the LED; alleviating the comparator from that job. Although the comparator spec says it can do +/-20mA, it's listed under Absolute Maximum ratings and I usually like to stay within 70% of those values. I also decided to switch to your original LT1711 because it does rail-to-rail output.

It seems to simulate perfectly, do you see any potential problems?

Thanks!

Edit
Quick follow-on question. I was reading the LT1711 datasheet and on page 8 under the section 'High Speed Design Techniques' it says to "keep source impedances as low as possible, preferable much less than 1kohm".

If I remember correctly, to calculate source impedance w.r.t the comparator, I need to short all voltage sources and calculate the equivalent resistance looking away from the comparator. That would make the inverting input have a source impedance 40K||10K == 8K and the non-inverting would be 10K.

To remedy this situation but still keep a 1V reference, I believe using a 1K for R1 and a 4K for R2 would do the trick. This gives me an equivalent impedance of 1K||4K = 800ohm. And for R3 I'll just simply use a 1K.

Is this all correct?
 

Attachments

Last edited:

Ron H

Joined Apr 14, 2005
7,063
Your calculations are correct.
In your case, I don't think your issue is speed. The input offset current on the LT1711 is as high as 3uA at 25°C. even with source resistances matched at 10k, this can give you an offset voltage of as much as 30mV. With them mismatched, the offset voltage can be higher.
Where do you get a 4k resistor? Do you buy 1% parts?
 

Thread Starter

SiegeX1

Joined Mar 10, 2009
25
I believe this application calls for both fast switching speed and a very good voltage reference. The LT1711 output will be disabling my dual MOSFET driver (in a push-pull configuration) during an overcurrent situation; sounds like I want the comparator to switch high ASAP.

As for the reference voltage, the current sensor puts out 2Vpp @ 5A full scale. However, I just realized that I didn't account for the fact that the real voltage to be measured will be 1Vpp as two 2Vpp outputs will be wire AND'd together through series resistors and only one output will be on at a time. Ideally, I would like to trip at 1.5A which to my calculation is a 300mV reference where each 1mV == 5mA.

I see that Linear has the LTC6650, an ultra precise 0.025% 2.5V regulator. This coupled with some 0.1% resistors (7.3K||1K) should give me a super solid 300mV reference. As you might have noticed, cost is not really a huge factor in this project (to some extent).

See attached revised schematic
 

Attachments

Ron H

Joined Apr 14, 2005
7,063
Do you really need that kind of precision? If it switches at 1.4A or 1.6A, would that be a problem?
What are you trying to protect?
 

Thread Starter

SiegeX1

Joined Mar 10, 2009
25
Hi Ron,

I'm trying to do exactly what is shown in Figure 3.6 with the disable setup shown in Figure 3.3. The two buffers driving the FETs in Fig 3.6 is the Si823x IC shown in Fig 3.3.

I'm trying to protect the FET's, the transformer and the load from a shoot-through situation.

As for precision, no, it probably doesn't matter if it's 1.4A vs 1.6A. But if the difference in cost is less than $15 then might as well go with the higher precision.
 

Attachments

Last edited:
Top