TRS202E: Can I power the V+ and V- rails using an external supply?

David,

I don't think I've ever seen this tried before but my assumption is it won't work. I say this because when I've used our RS232 devices with incorrect charge pump caps, I ended up seeing the device going into a 'lock out' mode where the device doesn't work on both the receiver side and the transmitting side. I think it requires the charge pumps to be active for the device to actually work.

David Stadille said:

It's a long story as to why I prefer to do this, glad to bore you with it if you like.

I'm always open to hearing a good story. 

Do you need a device where you can supply your own +12/-12V supply rails? You may need to look at this device instead: https://www.ti.com/product/GD65232

Otherwise, we don't have a device that can output +/-12V. 

-Bobby

Hi Bobby,

Thank you for the note.  Our application requires wide swings because the receivers that we will be sending data to have non-standard thresholds (they are a legacy homebrew sort of design, long story).  Additionally, we've noticed that using PC cards that implement RS-232 via charge-pump devices, the magnitude of the swings decreases with time (and we have to continuously plug and unplug the PCAs that have said homebrew receivers).   When we use the old school 75185 devices (for which we supply +/- 12V rails), the rails hold up and everything is fine.   So why not just use the 75185s you might ask?  In an upcoming ATE system, we will need 40 TX/RX pairs, full 15kV ESD, and even though I know the 75185 is pretty tough (and underrated) relative to ESD, I'd need 14 such devices and I am limited in terms of how much power is available.   The 75185 takes a fair amount of juice. 

One of the things that leads me to considering powering the device from external sources is the fact that the maximum ratings for V+ & V- are +/- 14V.  You can't get 14V from a capacitive doubler even at 6V input.  It seems like a subtle invitation to allow for external power. 

Thanks,

David 

Hi David,

Thanks for the background. The 75185 is probably going to act similar to the GD65232 device I suggested earlier. Not sure how much power savings you would get from using that suggestion. 

Clemens approach may be a better approach to this since you have a bit more control of the output impedance (which can help with the power) though with 40x T/Rs you'll likely still see a large dependency on power. I'm not really sure how to lower this without modifying the output impedance or the termination resistor to GND on the receivers. (Not having the charge pumps would help though as those just eat up current). This might be a situation where you look to increasing the power rail's output current capability.

-Bobby

Clemens, thank you very much for the suggestions here, I might in fact implement the system you have suggested.  But I also want to get to the bottom of the external power question, too. 

Bobby I'm still wondering if we just can't power the V+ and V- terminals externally, and again, this because the limits stated in the datasheet are beyond what the charge pump produces.  It seems like they left this option open.  Is there some old timer (I'm a half-million years old myself) there who could give us a reading?  

David,

Thanks for your patience here, Bobby is out of office today but will be back tomorrow. Please expect a response by 04/23/2025.

Regards,

Eric Hackett 

The V+/V− supplies are regulated, so in theory, the charge pumps should stop switching when the voltages are too high. But if there ever is one switching step, for whatever reason, then the voltage doubler will also work as a voltage halver in the other direction, and try to raise VCC to 6 V.

Hi David,

Unfortunately I'm probably the old timer for this device.

I'm 90% sure that the device won't work since it requires the charge pumps to be operational. the higher voltage on V+/V- may cause back biasing like Clemens pointed out as well which could damage the C1/C2 charge pump circuits.

Problem is this device went through a redesign on the die to make it in a 300mm wafer (old dies used 150mm and 200mm). So for future projects, you will probably start to get the new dies. We have an evaluation module (internal) for our 2T/2R RS232 devices. I will probably have to test this myself in the lab but it this may take a while as I think I need to get help from my marketing engineer to secure new die samples (probably wouldn't help if I tried to get samples myself and we end up with old dies which will be discontinued). 

Do you have a specific time frame when you need to know this? I'm worried it may take 2 weeks to secure samples. (assuming the new dies are in production now)....... 

-Bobby

Thanks for your thoughts Clemens :)

Bobby, 

Thank you for your advice on this.  OK, maybe I will re-think this.  I don't want you to go to a bunch of trouble.  This is for an ATE system we're building, and though we will ultimately test up to 320 units at once (in a series of cabinets), we may only build a few of these test stations over the next year or so.  I may go in the direction Clemens recommended early on.  Trying to get solid rails that don't sag with time, ESD, and a ton of TX/RX channels.  How about I feed back to you guys what I ultimately come up with? 

Dave

Hi Dave,

Sure. Please update us when you have more news.

If you change your mind and want me to look at trying to modify the device on an EVM, I can do this for you. Just requires time to get the correct samples.

-Bobby