27V input on ULN2001 is it safe ? 1

[LarsDK]

Hi,

I need advice in designing a relay board.

It is 24Volt relays. On the board today there are a diode over each relay. I have been thinking about change the diodes (7 * 1N4148) to a ULN2001.

But i am not sure is it ok to have a input 27Volt at the input on the ULN2001.

From PDF Absolute maximum ratings Vi = 30Volt

That do you say ? is it stupid to change the old board ?

Lars

 

[itsmoked]

Depends. If you have a functioning board doing the job now why spend 30secs changing anything. If you have to change things anyway for other reasons, I'd do it. The ULNs are robust chips.

Not clear on your 30V limit. The TI one is 50V.

Do check the chip power limit. You can not have all the outputs running near their individual limits at the same time. In fact not many channels at all.

Keith Cress
kcress -

http://www.flaminsystems.com

 

[LarsDK]

PCB must be changed because some new connectors.

But now when I have the file opened, there was an opportunity to optimize production time.

You know component count = time in production.

I have made a drawing to show that i mean.

The 50V i think that is on the output.

 

[itsmoked]

I C now.

Thanks for the nice drawing.

Yep.. NOPE! Don't bother.

1) The existing diodes will cost less than the ULN.

2) Time and again it has been proven over, and over, that assembly shops will not change their assembly charge over parts changes this inconsequential.
I have seen many cases where serious redesign effort was applied only to have the 'new' assembly quotes go up substantially. Using what you have evokes the assembler to 'charge the same'. Anything new invokes a 'new quote' that will cost more.
If you had some dorked-out connector that specifically drew complaints from the assemblers you might get a reduction, if, you specifically got it out of them before the redesign.

3) Your existing current loops between the relay coils and their individual diodes are much smaller than the ones that would result from the ULN design. Your new design could create more EMI.

4) Your existing system was designed to run the relays directly in whatever combination needed. Adding the ULN could restrict your system in a manner not present now. (see my previous comment)

5) Your existing system would be more reliable than with the addition of the ULN which adds another IC into the mean-time-between-failures equation.


Q.E.D. I would not do it.

Keith Cress
kcress -

http://www.flaminsystems.com

 

[IRstuff]

Just as a point of reference, I looked at all 4 datasheets available for this part at

http://www.datasheetcatalog.com/datasheets_pdf/U/L/N/2/ULN2001.shtml

And all 4 list output voltage rating of 50 V and input voltage rating of 30V.

But, you appear to be proposing adding essentially 4 additional components to the clamp diode that you're replacing. The IN4148 lists at 0.75 cents ea. in 100 quantities. Arrow lists the ULN2001 at 41 cents ea. in 2500 qty. I don't see much of a savings, unless there's some handling charge that your assembler charges.

TTFN

FAQ731-376

 

[benta]

Add to that that the ULN will invert the logic controlling the relays.

Benta.

 

[LarsDK]

Thank you all for your help.

Especially for Keith Cress of the convincing points.

So have decided to keep diodes.

Although the savings will be higher because we are hand fitting the components.

To IRstuff. My proposing is adding U1 and remove D1 to D7. The cost saving will be in labor ! not component price.

To Benta. No, it is not true. If you have any doubts then I would ask you to look at the drawing again.

Again, many thanks for your opinions. I hope that I may at some point return the help.

 

[IRstuff]

I understood the schematics. My point is that the ULN2001 approach replaces a 1N4148 with a clamp diode, 2 transistors and 2 resistors, i.e., 4 additional components, albeit, integrated into a single chip. This goes into what Keith inferred in his comments regarding reliability.

In addition to what Keith mentioned, any oddball fault on one channel can potentially affect the other channels in the ULN2001, given that they all share the same substrate, while your existing configuration has completely independent channels, at least, on the schematic shown.

TTFN

FAQ731-376

 

[LionelHutz]

Are you doing work on the board that connects to that connector? If so, then you could drop that board back to being just logic levels and remove the relay drive circuits from it. Then, you could justify doing what you propose.