Puzzling Pulse Transformer output 1

[swb1]

Hi,

I am pulsing a 1:1 pulse transformer with a 600 ns pulse and when I send 3 consecutive '1' the pulses on the output satrt to get smaller (about 350 ns). If I send 2 consecutive '1' the pulses on the output are fine. I tried putting a 1N4001 diode across the primary and secondary sides of the transformer and it didn't help. Can anyone offer any help??

Thanks,

swb1

 

[europus2002]

Hi swb1, what sort of transformer u have?? any datasheet?what type of circuit is it? .....This is the dV/dt limitation i guess....whats' yr transformer loading scenario? Can u reduce load on the output side? This will definitely help...

 

[swb1]

I am using a Pulse Engineering transformer PE-8272M. I have a 1N4001 diode connected across the primary and secondary sides. Cathode to pin 6 and anode to pin 4 on primary side and cathode to pin 1 and anode to pin 3 on the anode side. I also have a FET 2N3504 (drain) and 100 ohm resistor connected to pin 6 of the transformer.

Thanks,

swb1

 

[sreid]

If you're driving the primary unidirectionally, the third pulse may be saturating the core (DC buildup).

 

[swb1]

Not sure what you mean by unidirectionally?? How do I eliminate this DC buildup??

swb1

 

[MacGyverS2000]

My guess would be you're reaching the saturation point of the chosen core. Is your signal always 0 to 5V (or whatever max voltage you're using), or does it also swing negative (i.e., -5V)? If you're only going from 0 to positive and back to zero, I lean even harder towards saturation. A small schematic of the core and associated circuitry along wiht your signaling scheme would help.


Dan - Owner

http://www.Hi-TecDesigns.com

 

[europus2002]

Hi Swb1,
Looking at the datasheet, the primary inductance is 200uHenry which i feel is high for operating at 600nSec i.e 1.3MHz frequency.....Adversly speaking, the FET that u;r driving , may not be switching exactly 0-5 V. The Diodes 1N4001 used are good only upto few hundered Hz and would infact distort your signal if used in even KHz range. I would start debugging 1st by using
1) Use BAS16 or equivalent High speed Schottky diode
for achieving free wheeling at 1.3MHz across
transformer terminals instead of 1N4001 rectifier
diode.
If the above doesn't work...
2) Use High speed MOSFETs like IRF640 or so which has very low "ON resistance" less than 0.5Ohm and has atleast 3MHz switching capability
If this also doesn't work....
3) Last, i would suspect the transformer. If it has Iron core, use transformer that has Ferrite core. Also check the primary and secondary inductance they should be as low as possible of the order of 10's of microhenries...

R u using this transfomer for a specific reason??? Why not use a better Opto isolator??? Will solve all yr issues...Let us know what come out of this....thanks

 

[swb1]

I tried suggestion #1 and replaced the 1N4001 diode with a B0520WS (had it in house) schottky diode and there was no change. I am off to try suggestion #2.

swb1

 

[Skogsgurra]

Can we have standard English, please. These u and r may be fun for someone. But that doesn't work here.

Regarding the pulse shape: It is what magnetic cores do when you do not give them a DC-free input signal. A free-wheeling diode does not restore zero flux and will not help. Use a DC blocking capacitor in series with the primary to avoid DC saturation. And make sure that the transformer is designed for your intended use. There are specs like uVs and risetime and so forth that must be met in order to get good results.

Gunnar Englund

http://www.gke.org

 

[cbarn24050]

Hi, no cicuit presents a problem, I assume you are switching one end of the primary to ground with the other tied to supply. If thats the case then you need to remove the diodes. Use a fet with an avalance rated internal diode (irf series). Your core is saturating because the "off" volt seconds are smaller than the "on" volt seconds.

 

[swb1]

I used the DC blocking capacitor suggested by skogsgurra and it blocked it out alright The input signal to the transformer was blocked out completely, so that's definately not going to work. The signals coming out of the secondary side were like 30 ns wide.

swb1

 

[itsmoked]

Got a picture of your schematic?

http://www.nve.com/Downloads/il710t.pdf

Keith Cress
Flamin Systems, Inc.-

http://www.flaminsystems.com

 

[swb1]

How would I attach a picture of the schematic??

swb1

 

[itsmoked]

See: faq238-1161

It is very easy, takes less time, then reading the faq.

Keith Cress
Flamin Systems, Inc.-

http://www.flaminsystems.com

 

[swb1]

I don't have a .jpg of the schematic, Its done in schema.

swb1

 

[geekEE]

You can just do a screen capture and save it to JPG and use a paint program to edit out the extraneous stuff from the screen.

 

[europus2002]

Hi swb1,
Your computer keyboard should have "Prt Sc" (print screen) key....Open Photo editor . Open your schematic page, click on "Prt Sc" key and then open Photo editor and from Edit manu, click on "paste as new image"....you should get the .jpg of your schematic at saved location....

 

[Skogsgurra]

"The input signal to the transformer was blocked out completely, so that's definately not going to work"

Oh yes. It will work. But you have to size the capacitor accordingly. Not just pick one from the drawer.

Gunnar Englund

http://www.gke.org

 

[drpappy]

Can you increase the amplitude of the pulse? If so, the frequency response of the transformer may be a little better. Perhaps feed a signal generator into the transformer and see how it responds in the MHz range.

 

[drpappy]

I was expecting to get flamed for such a dumb post; so I'll do it myself: As already mentioned above in the first posts, increasing the voltage swing can make saturation worse, not better. This is regardless if it's a pulse transformer or not. Although what I did notice on the bench was that a transformer can saturate differently at several different frequencies like poles of a passive filter.