Winter Project #1

[Mike H]

Few more pix:

-

I was given a small drill press that fits my power drill so tried it out for the first time to make some more holes

-

-

-

-

[Mike H]

-

"Speaker protection module", Chinese kit, uses UPC1237 chip. Half the component values were changed after I finally found a decent datasheet for the IC, which shows e.g. how to precisely calculate the resistor values according to supply Voltage, and where you want the input threshold levels and so on.

I eventually managed to create a working LTspice model of the chip, not least aided by actually having one to look at and test (for example I discovered that the cap of the start-up delay timer RC network is discharged to zero whenever the chip is tripped to break the relay, and stays that way until the fault clears [input goes back to 0V]).

Subsequently I was able to produce the following schematic using LTspice – in this case the simulation was to demonstrate / test / tweak the overload input (pin P1). This pin responds to +0.72V, or >110 µA, but the one I've got needs 140 µA so the model had to be adjusted.

-

-

-

-

[Mike H]

Still need to put its feet on. I think I now know where they can be positioned.


Footnote: the case is really quite good, robust, thick panels; you get a solid aluminium skirted knob, power switch with solid aluminium button, Euro 3-pin fused mains inlet, 4 rubber feet, and more screws and nuts than you actually need, in case you lose any.

[Ray P]

I was given a small drill press that fits my power drill so tried it out for the first time to make some more holes

More holes! Lucky coincidence that they were in the right places for the sockets and stuff

[Mike H]

Six are provided, I needed four more in the (upper) middle

[Mike H]

A bit more info. This is already a couple of days old, meant to do some more today but keep being distracted.

Anyhoo.....

PW Texan / PE Rondo Power Amp Revisted

See also: http://www.angelfire.com/sd/paulkemble/sound8h.html
A Paul Kemble web page - PW (Practical Wireless) Texan amplifier and the PE Rondo.

I made the PE Rondo (where do they get these names from?) power amps (only)

Practical Electronics September 1973
CONSTRUCTIONAL PROJECTS
P.E. RONDO QUADRAPHONIC SOUND SYSTEM-1 by A. A. Cole
The first of a series of articles describing Quadraphonics in theory and practice with practical details of a decoder.

Practical Electronics October 1973
CONSTRUCTIONAL PROJECTS
P.E. RONDO QUADRAPHONIC SOUND SYSTEM-2 by A. A. Cole
Construction of pre-amplifiers, tone and balance controls and 20W amplifiers.

Who remembers quadaphonics? Was going to be the next new dogs dangly bits big must have. Vynil records encoded with an ultrasonic carrier for rear left and right speakers, over the top of the normal stereo groove. I kid you not. It sank without a trace of course.

Anyway, I just made a pair of the Rondo power amps for my own stereo integrated. The particular feature is that of an op-amp IC driving a p-p output stage, but because the op-amp supply Voltage rails have to be limited to a safe max. (+-15V), the output stage has to have some extra Voltage gain (as well as current gain) to to get a full swing between +-25V rails.


Move forward to 2019
Front-End as Operational Amplifier (with, as originally, limited Voltage supply rails)


Bit Of History:
The following is from Wireless World 1972, a "discretes op-amp" project, ulitimately comprising of a complete system as a phono pre, followed by a typical tone control block of the day (Baxandall treble and bass controls around an inverting op-amp). The author was saying that the op-amp chips at that time (1972) were too noisy for serious audio use, hence the idea of making your own, using discrete transistors and resistors of your preference. Which still sounds like an interesting option to play with.

I actually built a pair for stereo phono pre's; I remember having terrible trouble getting hold of the 4.7V Zeners, none of the usual suppliers seemed to have any of the sensible sized ones in stock so I ended up using huge, silver coloured, metal cased 5 Watt ones, or somesuch, which looked like overkill on the boards, but did the job.

One channel had a persistent fault in that after about half an hour or so after powering up, it would go silent and no sound would thereafter issue forth. At the time I never found out what was up, it was long time ago and I was still too green at this electronics lark.

Move forward 40 or so years, and now with the likes of LTspice to play with, it all becomes clear; there was a mistake on the published schematic and in the parts list, R2 should be 2.2k not 22k! At 22k the collector to emitter Voltage for Q4 is nearly at zero. Presumably in my original bad one, some thermal drift or somesuch eventually caused Q6 to be permanently on, Q1 unable to turn of off any more (is down to near zero collector current), so the output goes to + Volts and stays there. If memory serves that's what was happening, if I put a meter on the output, it's up near the positive supply rail and won't get off it. Unless you turn it off and leave it for a while.


-

Wireless World Discrete Op-Amp 1972 (corrected)

-

For some time I've been kicking around ideas for doing this using transistor arrays, mostly CA3046 – can't believe some of these spice files are dated 2010 & 2012.

The following is the first attempt at making one with a lower impedance output for the amplifier I'm making now, and using the CA3096 transistpr array IC (higher Vce and mixed npn & pnp available on same chip). Unfortunately the gain (Hfe) of the pnp's is pants, so the output doesn't work very well.

-

-


Works much better tho if the two pnp's are used together as a current mirror, and at much reduced current (not as low impedance emitter follower). This produced the following which works very well. For the first time in simulation, the output DC is near zero straightaway without having to "cheat" the value of a couple of resistors to make it so. I'm quite pleased with the result.

-

-

Note two array chips are used, transistors Qa1 – Qa5 is the CA3096, while Qb3 – Qb5 are the CA3046. An odd BC557 is added to do the output, as it has better gain. Another reason for choosing the positions is relative working Voltage, the CA3046 is the worse of the two, a minimum Vce of 15, it might be better, depending on what you get, but how would you know? Can't design for the possibilty that a chip you've got might be safe at a higher Voltage. Hence the CA3096 transistors are used for larger Voltage swings.

CA3046, 5 x npn
Vce, 15 min.
Hfe, 100 or better

CA3096, 3 x npn, 2 x pnp
npn Vce, 35
npn Hfe, 150 or better
pnp Vce, 40
pnp Hfe, 20 @ Ic = 1 mA (i.e., not very good!)
pnp Hfe, 40 @ Ic = 100 µA (so very low current use only, and still not marvellous)



That's all for now, next I was going to figure out how to wire up the CA3096 & CA3046 IC's together.

[Nick]

Who remembers quadaphonics

Me I am afraid, and I also remember those articles.

[Ant]

we did get the shibata stylus and microline profiles out of cd4 too, so perhaps it did leave a little bit of a legacy

[jack]

I thought that this rang bells - back in 2008, Jan Didden of Linear Audio fame, wrote a nice article on speaker protection for Elektor.

The article is downloadable for free from the Linear Audio site: https://linearaudio.nl/sites/linearaudi ... 050501.pdf

Interesting read.

[Mike H]

Thanks for that. Changed my mind about the overload input still need to work it out.

@ Ant: yes! Thanks.

[Mike H]

This is not forgotten, I think about it frequently, and how to finish it. Now, in my head, I think I know what the final circuit will look like. In my head.....