Vintage USRobotics External Modem & Power Supply

[tk3000]

Hello Folks,

Amongst the vintage computer hardware, parts, and accessories that are in my inventory is a couple of old analogue external modems using rs232 port (db9). I would like to test them, but I don't have their respective power supplies. USRobotics seemed to treat their power supplies and specs as an industrial secrete, so one can not find that basic information anywhere in the modem's case or documentation.

I believe that many of the usrobotics external modems used AC power supply instead of DC. So, basically the wall wart power adapter would simply step down the voltage and the modem itself would rectify it. It is not clear for me whether the ones I have use DC or AC. I tend to believe that using DC at the right voltage would not do any harm and probably would work out just fine; but I am not sure and don't know about the detail of these devices... Maybe 9VDC would be a good starting point?

The following are the aforementioned modems: USR5630G 56K V.92 Serial Controller Faxmodem (USR015630G, 64-245630-05R) AND USR5686E, USR325686E 56K V.92 External Faxmodem.

Believe it or not, they still have existing urls with all sort of information. The following are the urls:
1) https://www.usr-apr.com/support/s-prod-template.asp?loc=apnz&prod=5630g
2) https://www.usr.com/en-support/product/?prod=5686e

Any insight would be appreciated.

 

[Starfalcon]

I used one of those back in the day to go online nearly 20 years ago, I have it and the power supply stashed around here somewhere. Ill have to see if I can find it and I can let you know the specs on the wall wart.

 

[magnetik]

Not sure it helps but the adapter from my USR .34 dual standard is 20v but no polarity listed interestingly. So looks like an ac to ac adapter like you stated

 

[GiGaBiTe]

but no polarity listed interestingly.

The output tells you that it's 20v AC, 750 mAh.

I've had dozens of external modems over the years, and there was no real standard, even within the same brand. Most of the modems I had were in the 9-12V and could be AC or DC irritatingly enough.

But before you plug any of those wall warts in, CHECK THE VOLTAGE. Use a multimeter on the output, those old linear wall warts drift in voltage over time as the transformer windings degrade, and they can be far out of spec to the original voltage. I've had decades old wall warts that were double, triple and sometimes quadruple their original rated output because of winding degradation.

Before slamming in any old wall wart, I would take the case off and look at the power input circuitry. It should be fairly obvious if it wants an AC supply, because there's going to be a bridge rectifier, or half wave rectifier diode somewhere. If it just goes directly to some regulator, it wants a DC supply. As for the voltage, I'd start low and see what happens. 9v is pretty safe for those old modems, and if it doesn't behave, bump it up from there in 1-2v increments.

Additionally, if it's DC, make sure you have the correct polarity. Lots of things in the 90s used center negative barrels, where later on the standard switched to center positive, but both were in use at the same time. Going backwards polarity can cause the magic smoke to come out because lots of devices didn't have input polarity protection.

 

[jeffreyp]

Hello Folks,

Amongst the vintage computer hardware, parts, and accessories that are in my inventory is a couple of old analogue external modems using rs232 port (db9). I would like to test them, but I don't have their respective power supplies. USRobotics seemed to treat their power supplies and specs as an industrial secrete, so one can not find that basic information anywhere in the modem's case or documentation.

I believe that many of the usrobotics external modems used AC power supply instead of DC. So, basically the wall wart power adapter would simply step down the voltage and the modem itself would rectify it. It is not clear for me whether the ones I have use DC or AC. I tend to believe that using DC at the right voltage would not do any harm and probably would work out just fine; but I am not sure and don't know about the detail of these devices... Maybe 9VDC would be a good starting point?

The following are the aforementioned modems: USR5630G 56K V.92 Serial Controller Faxmodem (USR015630G, 64-245630-05R) AND USR5686E, USR325686E 56K V.92 External Faxmodem.

Believe it or not, they still have existing urls with all sort of information. The following are the urls:
1) https://www.usr-apr.com/support/s-prod-template.asp?loc=apnz&prod=5630g
2) https://www.usr.com/en-support/product/?

The output tells you that it's 20v AC, 750 mAh.

I've had dozens of external modems over the years, and there was no real standard, even within the same brand. Most of the modems I had were in the 9-12V and could be AC or DC irritatingly enough.

But before you plug any of those wall warts in, CHECK THE VOLTAGE. Use a multimeter on the output, those old linear wall warts drift in voltage over time as the transformer windings degrade, and they can be far out of spec to the original voltage. I've had decades old wall warts that were double, triple and sometimes quadruple their original rated output because of winding degradation.

Before slamming in any old wall wart, I would take the case off and look at the power input circuitry. It should be fairly obvious if it wants an AC supply, because there's going to be a bridge rectifier, or half wave rectifier diode somewhere. If it just goes directly to some regulator, it wants a DC supply. As for the voltage, I'd start low and see what happens. 9v is pretty safe for those old modems, and if it doesn't behave, bump it up from there in 1-2v increments.

Additionally, if it's DC, make sure you have the correct polarity. Lots of things in the 90s used center negative barrels, where later on the standard switched to center positive, but both were in use at the same time. Going backwards polarity can cause the magic smoke to come out because lots of devices didn't have input polarity protection.

I would a few USR modems through the early years. 28.8k baud was the fastest.
The modems came with a plug-in AC wall to DC converter. Nothing runs on AC.

 

[Nenu]

I have a functional USRobotics 56K modem here somewhere.
I know its power supply is 9V AC, I accidentally used it on one of my guitar pedals that requires 9V DC and blew it!
When I find it later I'll check the polarity and exact model.

ps
Use the wrong PSU and it probably wont survive, dont just experiment.


Edit
I found 2 USR 56K modems.
They dont appear to have model numbers written on them, only serials.

The one I know takes 9V AC is the Sportster Flash but I cant locate its power supply atm, I can visualise it in my head.

The other is labelled 56K Faxmodem, its PSU is 9V AC (from markings on it) but is dead now so cannot check polarity.
I was given this but have never used it.

 

[Nenu]

I would a few USR modems through the early years. 28.8k baud was the fastest.
The modems came with a plug-in AC wall to DC converter. Nothing runs on AC.

Not true.

 

[GiGaBiTe]

Nothing runs on AC.

False.

Lots of old computers and electronics used AC wall warts and internally converted it to DC. Examples would be Commodore machines and the TI-99/4a. There were plenty of external modems that used AC wall warts. I have a Hayes modem that uses one. Even the USR modem in the post above proves you wrong.

 

[Zepher]

False.

Lots of old computers and electronics used AC wall warts and internally converted it to DC. Examples would be Commodore machines and the TI-99/4a. There were plenty of external modems that used AC wall warts. I have a Hayes modem that uses one. Even the USR modem in the post above proves you wrong.

technically he is right though, even though the device does the rectifying, it runs on DC.

 

[Nenu]

technically he is right though, even though the device does the rectifying, it runs on DC.

Being further pedantic, he is wrong
ie Cookers, lights, heaters, motors, passive speakers, electrocution chairs ...

 

[jeffreyp]

False.

Lots of old computers and electronics used AC wall warts and internally converted it to DC. Examples would be Commodore machines and the TI-99/4a. There were plenty of external modems that used AC wall warts. I have a Hayes modem that uses one. Even the USR modem in the post above proves you wrong.

Any electronic device whether it has internal or externally powered has to have a full wave rectifier and assorted components.

 

[jeffreyp]

Not true.

Really? https://stason.org/TULARC/pc/modems...ION-Fax-Modem-asynchronous-US-ROBOTIC-32.html

 

[jeffreyp]

Not true.

Yes true.

 

[Nenu]

Really? https://stason.org/TULARC/pc/modems...ION-Fax-Modem-asynchronous-US-ROBOTIC-32.html

Yes true.

This is your post I quoted:

I would a few USR modems through the early years. 28.8k baud was the fastest.
The modems came with a plug-in AC wall to DC converter. Nothing runs on AC.

You said nothing runs on AC.
Incorrect and proven.


Also the modem you just quoted doesnt conform to your post either

The modems came with a plug-in AC wall to DC converter.

Thats an internal modem, it cannot use "a plug-in AC wall to DC converter"

 

[jeffreyp]

This is your post I quoted:


You said nothing runs on AC.
Incorrect and proven.

This is your post I quoted:


You said nothing runs on AC.
Incorrect and proven.


Also the modem you just quoted doesnt conform to your post either

Thats an internal modem, it cannot use "a plug-in AC wall to DC converter"

I said not without at least a bridge rectifier and assorted components. Which converts AC to DC. No modern electronic component that I know of runs on a dame sine-wave. LoL!!

 

[Nenu]

I said not without at least a bridge rectifier and assorted components. Which converts AC to DC. No modern electronic component that I know of runs on a dame sine-wave. LoL!!

That was you 'later' trying to avoid embarrassment.

The device either accepts AC or DC from its external user inputs.
That defines what it runs on.
I'm an electronic engineer btw

 

[jeffreyp]

That was you 'later' trying to avoid embarrassment.

The device either accepts AC or DC from its external user inputs.
That defines what it runs on.
I'm an electronic engineer btw

I am a retirered Electronics design engineer. I desinged radiatiation detectors. 34 years at SLAC Stanford Linear Accelerator Center. btw...

 

[Nenu]

I am a retirered Electronics design engineer. I desinged radiatiation detectors. 34 years at SLAC Stanford Linear Accelerator Center. btw...

You should have known better then

Good projects you were involved in, kudos.

 

[jeffreyp]

You should have known better then

Good projects you were involved in, kudos.

Thanks for the advice fathead. I said nothing wrong.

 

[Nenu]

Thanks for the advice fathead. I said nothing wrong.

lol

 

[tk3000]

Not sure it helps but the adapter from my USR .34 dual standard is 20v but no polarity listed interestingly. So looks like an ac to ac adapter like you stated

Thanks for the pics!

Since mine may use AC as well and AC is given as a RMS value - which then is numerically roughly equivalent the DC value - I am assuming that a slightly lower or equivalent DC voltage is not likely to cause any damage or faulty operation. I was trying it with 9VDC and the modem powers on (power led on), the other activity leds seem to blink when I try to communicate with the modem via a terminal program (putty) but there is no response from the modem (whenever I issue the few AT commands that I can remember [ATI[number], ATA, etc). Subsequently, I tried it with 12VDC to no avail either.

 

[tk3000]

Before slamming

The output tells you that it's 20v AC, 750 mAh.

I've had dozens of external modems over the years, and there was no real standard, even within the same brand. Most of the modems I had were in the 9-12V and could be AC or DC irritatingly enough.

But before you plug any of those wall warts in, CHECK THE VOLTAGE. Use a multimeter on the output, those old linear wall warts drift in voltage over time as the transformer windings degrade, and they can be far out of spec to the original voltage. I've had decades old wall warts that were double, triple and sometimes quadruple their original rated output because of winding degradation.

Before slamming in any old wall wart, I would take the case off and look at the power input circuitry. It should be fairly obvious if it wants an AC supply, because there's going to be a bridge rectifier, or half wave rectifier diode somewhere. If it just goes directly to some regulator, it wants a DC supply. As for the voltage, I'd start low and see what happens. 9v is pretty safe for those old modems, and if it doesn't behave, bump it up from there in 1-2v increments.

Additionally, if it's DC, make sure you have the correct polarity. Lots of things in the 90s used center negative barrels, where later on the standard switched to center positive, but both were in use at the same time. Going backwards polarity can cause the magic smoke to come out because lots of devices didn't have input polarity protection.

Yeah, I have some variable dc power supplies that are relatively new and in good working order. I have tried 9VDC and 12VDC to no avail. The modem powers on, leds lights up when I try to communicate with the modem via a terminal program via serial interface; but there is no response. The only type of response I managed to get was some noise once I entered the ATA command in the terminal emulator (putty).

Having said that, I am not sure about all the configuration details, etc; it has been over 20 years since last time I used an analogue modem. Also, the vast majority of the modems I had during the heyday of dialup were internal ones.

Thanks for the hints on the voltage and polarity. I was taking the path of less effort at first. It would be possible for the polarity in the power jack to be reversed, but that would be unlikely (at least based on my fading distant memory). And, I would imagine that a reputable manufacturer like usrobotics probably would not overtly cut costs to a point of avoiding using a reverse polarity diode. Yeah, but it is all in the realm of possibilities; and it is better to be safe than sorry.

 

[toast0]

If it's making answer noises when you do ATA, it might be your serial port settings or your cable? Also try ATE1 to enable local echo? and/or AT&F0 to enable sensible defaults?

 

[tk3000]

I have a functional USRobotics 56K modem here somewhere.
I know its power supply is 9V AC, I accidentally used it on one of my guitar pedals that requires 9V DC and blew it!
When I find it later I'll check the polarity and exact model.

ps
Use the wrong PSU and it probably wont survive, dont just experiment.


Edit
I found 2 USR 56K modems.
They dont appear to have model numbers written on them, only serials.

The one I know takes 9V AC is the Sportster Flash but I cant locate its power supply atm, I can visualise it in my head.

The other is labelled 56K Faxmodem, its PSU is 9V AC (from markings on it) but is dead now so cannot check polarity.
I was given this but have never used it.

9Volts sounds reasonable for those type of devices. But, to my knowledge, I don't have any wall wart that output AC; all the ones I have output DC given that electronic devices pretty much always use DC due to its nature.

 

[GiGaBiTe]

Any electronic device whether it has internal or externally powered has to have a full wave rectifier and assorted components.

Nope. There are plenty of electronics that run directly off the mains AC, or use half-wave rectification.

One common example are LED bulbs. To make them only cost a couple of dollars, they've cut the component count down to almost nothing. Most cheap LED bulbs use capacitor droppers, and sometimes even just resistors. This puts a DC component, a light emitting diode, on full mains voltage. And it works because once you get enough of them in series, the voltage drop is close enough to the mains voltage that it works. Not to say they'll have a long life, because they're smashed with so much current to get them as bright as possible that they don't last.

SMPS supplies are another. How do you think the SMPS supply starts? There's a kickstart circuit responsible for getting the mosfets going in some designs, and those often have no more than a single diode, a few resistors and a capacitor to work.

 

[tk3000]

nd/or AT&F0 t

Hardware wise, I am using the proper cables (tried two different straight-through serial cables) and tried a native serial port and a usb-rs232 bridge converter (prolific 2303GS). Also, tried two different computers with the same results. It would be too little luck for both cables and both rs232 ports to be failing with identical outcomes. So, likely the modem is not getting the proper power, it is defective, or it is software configuration.

Accessing the "Modem Properties" in the device manager of windows 10, the status shows "The device is working properly". But when I go to the Diagnostics tab and query the modem there it show an error: "The modem failed to respond. Make sure it is properly connected..."

The configuration used for the serial port is: 115200, Data bits=8, Parity=None ,Stop Bits=1,Flow Control=None". The modem maximum speed is set to 115200. I don't recall whether or not the serial port in the modem and computer would negotiate a different speed if necessary. Tried some variations, such as changing the flow control to "Hardware" but it did not change a thing.

Also tried your suggestions (ATE2 and AT&F0) and still no results.

 

[Starfalcon]

I am still looking for my modem and wall wart, I know its here somewhere...just a lot of boxes to look through for it. Hopefully I can get you the info your looking for.

 

[Nenu]

9Volts sounds reasonable for those type of devices. But, to my knowledge, I don't have any wall wart that output AC; all the ones I have output DC given that electronic devices pretty much always use DC due to its nature.

Assuming its an AC device, you will need a 9V AC adapter if you dont want to risk damage.
BUT that assumes its an AC device. Putting an equivalent voltage AC adapter on a DC device will probably kill it.
As my Boss Metal Zone MT-2 guitar pedal is aware, expensive mistake!


Overvolting a DC device with DC can also kill it easily. This is because the internal effective resistance (R) remains the same and Power = (Voltage squared)/R (assuming it has poor voltage regulation as is common in older devices).
A small change in voltage can give a large increase in power use AND heat.
ie
Lets 'assume' the device has 10 ohm Resistance, in this case:
Power = Voltage squared/10

With the correct 9V PSU
Power = 9x9/10 = 49/10 = 4.9W

With 12V PSU
Power = 12x12/10 = 144/10 = 14.4W

You can see thats 3x the power just from what appears a small change in voltage!
Regardless of the device Resistance, the above law will apply the same with the voltages shown above.
So beware in case its not an AC model.


You could open it up to see if there is a rectifier chip, or 4 diodes used as a bridge rectifier, or 1 or 2 diodes used as a 1/2 wave (or 2 diode full wave) rectifier.
If there is any type of rectifier it needs AC input.
If there is no rectifier anywhere its DC input.

 

[toast0]

Hardware wise, I am using the proper cables (tried two different straight-through serial cables) and tried a native serial port and a usb-rs232 bridge converter (prolific 2303GS). Also, tried two different computers with the same results. It would be too little luck for both cables and both rs232 ports to be failing with identical outcomes. So, likely the modem is not getting the proper power, it is defective, or it is software configuration.

Accessing the "Modem Properties" in the device manager of windows 10, the status shows "The device is working properly". But when I go to the Diagnostics tab and query the modem there it show an error: "The modem failed to respond. Make sure it is properly connected..."
The configuration used for the serial port is: 115200, Data bits=8, Parity=None ,Stop Bits=1,Flow Control=None". The modem maximum speed is set to 115200. I don't recall whether or not the serial port in the modem and computer would negotiate a different speed if necessary. Tried some variations, such as changing the flow control to "Hardware" but it did not change a thing.

Also tried your suggestions (ATE2 and AT&F0) and still no results.

You might try reducing the data rate and see if that helps? Most modems above say 2400 baud should have enough smarts to autodetect the serial bitrate from the AT string, but may default to 9600. But sounds like it's definitely not happy. Could contact USR and ask them ... they apparently packed the wrong adapters for a while, so they might have information available?

 

[tk3000]

Assuming its an AC device, you will need a 9V AC adapter if you dont want to risk damage.
BUT that assumes its an AC device. Putting an equivalent voltage AC adapter on a DC device will probably kill it.
As my Boss Metal Zone MT-2 guitar pedal is aware, expensive mistake!


Overvolting a DC device with DC can also kill it easily. This is because the internal effective resistance (R) remains the same and Power = (Voltage squared)/R (assuming it has poor voltage regulation as is common in older devices).
A small change in voltage can give a large increase in power use AND heat.
ie
Lets 'assume' the device has 10 ohm Resistance, in this case:
Power = Voltage squared/10

With the correct 9V PSU
Power = 9x9/10 = 49/10 = 4.9W

With 12V PSU
Power = 12x12/10 = 144/10 = 14.4W

You can see thats 3x the power just from what appears a small change in voltage!
Regardless of the device Resistance, the above law will apply the same with the voltages shown above.
So beware in case its not an AC model.


You could open it up to see if there is a rectifier chip, or 4 diodes used as a bridge rectifier, or 1 or 2 diodes used as a 1/2 wave (or 2 diode full wave) rectifier.
If there is any type of rectifier it needs AC input.
If there is no rectifier anywhere its DC input.

As your probably imagine, before posting in this forum I tried to contact usrobotics in order to obtain the power specs for the modem. Haven't heard from them in a few days, so I assumed that they would not response (not many companies provide support for ancient hardware). To my surprise though, I received an answer today for one of my external modems (I have two as indicated in the first post of this thread). The following is the answer:
"
Thank you for choosing USRobotics.
The USR325686E uses a 9V AC 1A power adapter.
I hope this helps.
"
But, anyway, I went ahead and opened one of the modems (only two screws, the other modem does not seem to use screws but only plastic tabs so I did not open it yet), the USR325686E. And, indeed, it has a linear voltage regulator (7805CD2T) and a full bridge rectifier in one smd chip (DF04S).

I imagine that the DC current would pass straight through the bridge rectifier, and the circuit would work as intended. Unless, some part of the circuit is using AC. And I understand that landlines uses DC, except when ringing when it superimposes an AC signal over the DC; so I don't know whether or not AC pays any role in this modem's circuitry. Below is a pic of the modem's circuit board:

 

[tk3000]

You might try reducing the data rate and see if that helps? Most modems above say 2400 baud should have enough smarts to autodetect the serial bitrate from the AT string, but may default to 9600. But sounds like it's definitely not happy. Could contact USR and ask them ... they apparently packed the wrong adapters for a while, so they might have information available?

I did not know about usrobotics shipping modems with the wrong power adapters. Anyhow, I did contact them few days ago; but did not hear back from them until today. Today, I got a response from them:
"
Thank you for choosing USRobotics.
The USR325686E uses a 9V AC 1A power adapter.
I hope this helps.
"
I will try a lower baud rate.

Thanks.

 

[Nenu]

Cool, you know how to proceed.
Fingers crossed nothing is damaged.