Rather than solder wire from the PCB to the heatsink mounted power transistors, is there a better (more aesthetically pleasing) way?
I noted in some photos that some appear to use a connector of sorts - possibly a 'female/female' solder socket? - with a covering sleeve.
Suggestions (and where to purchase if available) appreciated.
Thanks.
I noted in some photos that some appear to use a connector of sorts - possibly a 'female/female' solder socket? - with a covering sleeve.
Suggestions (and where to purchase if available) appreciated.
Thanks.
I think you can get "sockets" for all of the currently popular power transistor packages, though I only recall seeing such things actually used with TO-3 packages. (Another superannuated guy will probably tell me they never made a socket for the TO-36 package.) You may consider them a more aesthetic choice for the transistor end of the wire. On the PWB end, there must be dozens, perhaps hundreds, of choices among standard pin-header, Berg, and Molex connectors. Personal computer wiring harnesses and internal power cabling may be a ready source of pre-assembled cables.
Another possibility on the PWB end would be push-on spade terminals, sometimes called flag terminals. They have a human-factors problem of correctly connecting 6 (or more) individual loose wires, all with the same terminal type.
I would give careful, critical, thought before using any of these ideas. Connections to power output devices can be expected to carry several amps, perhaps peak currents of tens of amps. IR drop across a connector can adversely affect performance. Vibration and thermal cycling can work connections loose. If I wanted to go for a sexy appearance, I'd consider soldering wires to the transistor leads, and make connections on the PWB end with a barrier block, and fork or ring terminals. Using a fanning strip with the fork terminals can help alleviate the problem of incorrect connections.
Dale
Another possibility on the PWB end would be push-on spade terminals, sometimes called flag terminals. They have a human-factors problem of correctly connecting 6 (or more) individual loose wires, all with the same terminal type.
I would give careful, critical, thought before using any of these ideas. Connections to power output devices can be expected to carry several amps, perhaps peak currents of tens of amps. IR drop across a connector can adversely affect performance. Vibration and thermal cycling can work connections loose. If I wanted to go for a sexy appearance, I'd consider soldering wires to the transistor leads, and make connections on the PWB end with a barrier block, and fork or ring terminals. Using a fanning strip with the fork terminals can help alleviate the problem of incorrect connections.
Dale
Dale,
Thanks very much for your reply.
I'd seen some pics on the forum which looked neat & tidy and seemed to have some form of connector involved between cable and transistor.
I'd bought some from an electronics store long ago but guess they were just loose connector pins from some or other socket connectors they stocked. Did the job.
Your point on the thermal cycling is well made.
Guess I'll just solder direct and use heat shrink tubing or such to cover the joints 🙂
Thanks very much for your reply.
I'd seen some pics on the forum which looked neat & tidy and seemed to have some form of connector involved between cable and transistor.
I'd bought some from an electronics store long ago but guess they were just loose connector pins from some or other socket connectors they stocked. Did the job.
Your point on the thermal cycling is well made.
Guess I'll just solder direct and use heat shrink tubing or such to cover the joints 🙂
Lengthening connections to modern power devices can cause RF stability trouble.
I would also be wary about one of the connections going intermittent, putting dc on the output to your speakers. These wires are critical.
I would also be wary about one of the connections going intermittent, putting dc on the output to your speakers. These wires are critical.
For 3pin transistors (BJTs and mosFETs), do any of the 3 wires need to be close coupled?
Do any of the wires need to be separated?
For a pair of complementary output devices that have a 6wire connection, do any of the rules for the above 3pin change?
Do any of the wires need to be separated?
For a pair of complementary output devices that have a 6wire connection, do any of the rules for the above 3pin change?
With ~100V and 10A class AB bursts around the wires had better be well insulated, thick enough and avoiding large loop areas
Thanks for all the replies
As a related question, if the O/P transistors are far enough away from the PCB should one 'braid' the ECB wires or leave them as a 'straight run' ('not twisted')?
As a related question, if the O/P transistors are far enough away from the PCB should one 'braid' the ECB wires or leave them as a 'straight run' ('not twisted')?
I would loose twist E and C and keep B close. Magnetic coupling from the class B output wire loops can be the cause of unexpectedly high distortion when it reaches into other circuits
Here's a guy who is using a kind of screw terminal block directly on transistor leads. You can see the PWB assembled without output transistors in the first photo on his main page, at:
http://www.renardson-audio.com/index.html
Another view ( http://www.renardson-audio.com/trans-add3.jpg ) :
With output transistors attached ( http://www.renardson-audio.com/final2.jpg ) :
And you can find quite a few other photos - as well as discussion about the pros and cons of this approach - by poking around in his web pages.
The same concept could probably be used with the classic "barrier block" style screw terminal strips if you crimp a fork- or ring-terminal onto the transistor leads. That may give a more "heavy duty" or "industrial" appearance to the construction, but I rather like the looks of these more modern screw terminal blocks.
Dale
- Status
- Not open for further replies.