02-08-2014, 06:38 PM
A good deal of effort (time and money) went into the design and development of this project. The fundamental problem with designing an adjustable and regulated all-semiconductor HT PSU is one of selecting suitable semiconductors that have a sufficiently high voltage rating and a reasonable d.c. current gain at the same time. Usually, high voltage transistors have a relatively low hFE. This can usually be overcome by using Darlington pairs (or even triples), but things then get complicated: I like to keep my designs as simple as possible, provided that they meet the essential specification requirement. This design does that.
That spec. is for an adjustable O/P from 100 to 300 v.d.c., regulated, up to 100 mA output current, short-circuit proof (current foldback) and with a low noise content on the O/P voltage. The HT O/P is floating with respect to ground / chassis. (Useful for things like the R1155, etc.). Plus 6.3 v.a.c. at 5 amps. Monitoring of d.c. O/P voltage and current is essential, of course, as is the ability to see what that O/P voltage actually is, prior to connecting it to the external load. Plus the ability for possible future maintenance without too much dismantling - and all inside a metal case of a sensible size. This PSU meets that specification.
A few comments on the design seem worthwhile. These will be useful should anyone copy this cct. and build one of their own.
► A choke-input filter was chosen not so much to use the improved regulation of that cct. over the usual C-L-C filter, but to keep its output voltage down (with the given transformer) to a level commensurate with the voltage ratings of the transistors used. The chosen ohmic value of the bleeder resistor helps in that respect. The 2N5416 has a Vce of 300 v. and for the 2N3439, that is 350 v. The ratings for the IRF340 are more than adequate for the voltages and currents associated with it. There is no specific reason for the particular transformer / rectifier / choke combination as chosen. The two conditions that need to be met are maximum N/L voltage and required current capacity. And in that regard, a choke with a higher current rating (and at 10 Henries) should provide for a greater O/P current capability.
► The feedback control loop is fairly typical, but note that the use of a FET for the series-pass device provides for an 'all-voltage' design of this part. The usual requirement with a bipolar transistor in this position - requiring high Vce and simultaneously a high hFE - is thus removed. (Initial attempts along that classical line gave many design problems).
► The current-limiting is of the foldback type: as soon as the O/P current exceeds about 120 mA, the O/P voltage is progressively reduced as that current increases - and in the event of a dead short cct., the 100 mA fast-blow fuse blows. In practice, when powering something like a wireless with a capacitor of substantial capacity across its HT rail, it might be thought that that capacitor would 'trip' that current overload cct. However on test, I found that connecting a fully-discharged 47 µF cap. across the O/P voltage terminals did not cause a problem.
► Although a DPDT switch is used to connect / disconnect the HT O/P voltage to the external load, only one pole of that switch is used: this was considered to be adequate. (If I had had a 3-pole switch, I would have used that). The other pole is used to switch a yellow neon bulb to indicate that the HT is on. This gives a clear warning that HV HT is present at the O/P terminals and thus the load. Accidentally putting 300 v.d.c. across your chest is something to be avoided! The position of that neon in the cct. ensures that the intensity of the glow of that neon is independent of the magnitude of the O/P voltage.
► The 150 pF capacitor, connected B to C of the 2N34329, was found to be an essential fitment to ensure h.f. stability. Without it, there was the tendency for that stage to oscillate and destroy most of the semiconductors! And those semiconductors are not exactly cheap, either.
► The noise & hum level at the O/P is well into the 'units of mV' range - except, of course, when the over-current cct. kicks in.
► The 2N5416 is a simple constant-current source which help to ensure that the gain of the voltage comparator cct. (built around the 2N3439) remains sufficiently high. This ensures a good voltage-regulation characteristic.
► The case I used just simply happened to be available. All of the structural metalwork was fabricated from salvage scrap.
And finally, I am not a professional analogue cct. designer! So if anyone can see any short-comings in this design, please speak. All I can say is that this cct. has undergone substantial 'soak-testing' without any mishap or malfunction.
Photos. are below; cct. diag. will be added as an additional post. The quality of the photo. of the control board is poor since it was necessary to substantially reduce its resolution so that it met the upload requirement.
[attachment=11023] [attachment=11024] [attachment=11025]
[attachment=11026] [attachment=11027]
Al. / Aug. 2, '14 //
That spec. is for an adjustable O/P from 100 to 300 v.d.c., regulated, up to 100 mA output current, short-circuit proof (current foldback) and with a low noise content on the O/P voltage. The HT O/P is floating with respect to ground / chassis. (Useful for things like the R1155, etc.). Plus 6.3 v.a.c. at 5 amps. Monitoring of d.c. O/P voltage and current is essential, of course, as is the ability to see what that O/P voltage actually is, prior to connecting it to the external load. Plus the ability for possible future maintenance without too much dismantling - and all inside a metal case of a sensible size. This PSU meets that specification.
A few comments on the design seem worthwhile. These will be useful should anyone copy this cct. and build one of their own.
► A choke-input filter was chosen not so much to use the improved regulation of that cct. over the usual C-L-C filter, but to keep its output voltage down (with the given transformer) to a level commensurate with the voltage ratings of the transistors used. The chosen ohmic value of the bleeder resistor helps in that respect. The 2N5416 has a Vce of 300 v. and for the 2N3439, that is 350 v. The ratings for the IRF340 are more than adequate for the voltages and currents associated with it. There is no specific reason for the particular transformer / rectifier / choke combination as chosen. The two conditions that need to be met are maximum N/L voltage and required current capacity. And in that regard, a choke with a higher current rating (and at 10 Henries) should provide for a greater O/P current capability.
► The feedback control loop is fairly typical, but note that the use of a FET for the series-pass device provides for an 'all-voltage' design of this part. The usual requirement with a bipolar transistor in this position - requiring high Vce and simultaneously a high hFE - is thus removed. (Initial attempts along that classical line gave many design problems).
► The current-limiting is of the foldback type: as soon as the O/P current exceeds about 120 mA, the O/P voltage is progressively reduced as that current increases - and in the event of a dead short cct., the 100 mA fast-blow fuse blows. In practice, when powering something like a wireless with a capacitor of substantial capacity across its HT rail, it might be thought that that capacitor would 'trip' that current overload cct. However on test, I found that connecting a fully-discharged 47 µF cap. across the O/P voltage terminals did not cause a problem.
► Although a DPDT switch is used to connect / disconnect the HT O/P voltage to the external load, only one pole of that switch is used: this was considered to be adequate. (If I had had a 3-pole switch, I would have used that). The other pole is used to switch a yellow neon bulb to indicate that the HT is on. This gives a clear warning that HV HT is present at the O/P terminals and thus the load. Accidentally putting 300 v.d.c. across your chest is something to be avoided! The position of that neon in the cct. ensures that the intensity of the glow of that neon is independent of the magnitude of the O/P voltage.
► The 150 pF capacitor, connected B to C of the 2N34329, was found to be an essential fitment to ensure h.f. stability. Without it, there was the tendency for that stage to oscillate and destroy most of the semiconductors! And those semiconductors are not exactly cheap, either.
► The noise & hum level at the O/P is well into the 'units of mV' range - except, of course, when the over-current cct. kicks in.
► The 2N5416 is a simple constant-current source which help to ensure that the gain of the voltage comparator cct. (built around the 2N3439) remains sufficiently high. This ensures a good voltage-regulation characteristic.
► The case I used just simply happened to be available. All of the structural metalwork was fabricated from salvage scrap.
And finally, I am not a professional analogue cct. designer! So if anyone can see any short-comings in this design, please speak. All I can say is that this cct. has undergone substantial 'soak-testing' without any mishap or malfunction.
Photos. are below; cct. diag. will be added as an additional post. The quality of the photo. of the control board is poor since it was necessary to substantially reduce its resolution so that it met the upload requirement.
[attachment=11023] [attachment=11024] [attachment=11025]
[attachment=11026] [attachment=11027]
Al. / Aug. 2, '14 //