Troubleshooting Telephone Dialer Misdialed Numbers - Holtech Brand

[Phrak]

I have a friend with a telephone dialer (brand name Holtech) in his design. Some numbers are getting misdialed. We suspect the line interface circuit.

Does anyone have topology and component values for the transformer and other components typically used? It has to meet the interface requirements.

 

[berkeman]

I may have that in a book at home. I'll make a note to check when I get home from work.

Are you able to see a pattern in the misdials? Do you have access to analyzers that can help you see what is wrong? Does the autodialer have an option to slow down its dialing (the length of each DTMF tone), and does that help?

 

[Phrak]

I may have that in a book at home. I'll make a note to check when I get home from work.

Are you able to see a pattern in the misdials? Do you have access to analyzers that can help you see what is wrong? Does the autodialer have an option to slow down its dialing (the length of each DTMF tone), and does that help?

The dialer is a single chip that takes a oscillator and interface. I'm afraid I don't know a lot of details, but it's an integrated package that looks for dial tone, then dials, then waits for either pick-up or busy. I believe pick-up might be signaled by a ~16 KHz burst, then throughputs a recoded message.

I don't know about a pattern. I'd like to see a scope on it too, but it's 70 miles away. No frequency analyser is available.

I think the interface design involved too much guess-work, and not enough research.

 

[berkeman]

The dialer is a single chip that takes a oscillator and interface. I'm afraid I don't know a lot of details. I'd like to see a scope on it too, but it's 70 miles away. No frequency analyser is available.

I think the interface involved too much guess-work, and not enough research.

Would you be able to post a list of some of the misdial patterns?

 

[Phrak]

Would you be able to post a list of some of the misdial patterns?

No, sorry. I'll see If I can get some.

I found the Holtek (corrected spelling) website though. http://www.holtek.com.tw/english/products/default.htm"

I'm going to have to get the part number. Otherwise we're going on too much iffy information.

 

[berkeman]

Brought my old phone electronics book into work:

https://www.amazon.com/dp/B000RRHZCW/?tag=pfamazon01-20

Brings back memories flipping through the pages!

 

[Phrak]

Brought my old phone electronics book into work:

https://www.amazon.com/dp/B000RRHZCW/?tag=pfamazon01-20

Brings back memories flipping through the pages!

You and my friend might have some things to talk about. I believe a telephone circuit was his first design. It was something on the receive-side involving way too many wire-wrapped PLLs and other stuff sometime in the 1980s. This current design is 7x4" of realestate! Good grief. The reason, from what I gather, is that ICs for these sort of circuits have come and gone as OMC devices, so it's built-up from pieces involving the dialer, 5 PLLs and 3 DPST relays to switch various transmitters and receivers in and out.

I'm afraid I'm only getting really sketchy information on this problem. The troubleshooting data base consists of calling one long distance cell phone and one local land-line. The dialer connects to the local land line OK. Dialing the cell phone returns a busy signal. The busy signal is real.

Stay tuned for more if you can take it.

 

[berkeman]

Interesting. What about cell phones on other carriers? Verizon versus AT&T versus whoever... I once had a problem accessing my work voicemail mailbox using my cell phone -- the password I had on the mailbox included a "1" digit, and for some reason, that was not making it through. Changing the password to avoid the "1" fixed my access problem. (A very experienced IT person at my work told me of the 1-problem and cell phones)

There must be a standard quality measure (frequency tolerances and SNRs) for DTMF signalling. Maybe even some test instrumentation that could be leased or borrowed. Maybe some test labs have the equipment, and could test the device to give your friend a report that could help him debug the problem. Yeah, look for telcom testing labs... I think there are several here in Silicon Valley.

 

[Phrak]

I'll let he know about the services. That's weird with the '1' access number. I gave him my wife's Verizon number to call. Don't know if he tried it or not. Either that or it didn't work either. I'm going to recommend sending tones over a longer time span to give the receiving circuits at the exchange longer to lock-on. It's worth a try.

He's using the HT9320 IC, btw, http://www.holtek.com.tw/english/docum/comm/9320.htm"

The datasheet quotes a tone error of 0.74% of the worst case generated frequency.