Good to see you on here Bob. Been awhile. Yeah the 723 design would be great. But I have a bunch of boards made up using the 7808 and transistor design. Solder together and go. For me anyways.
-
You can now help support WorldwideDX when you shop on Amazon at no additional cost to you! Simply follow this Shop on Amazon link first and a portion of any purchase is sent to WorldwideDX to help with site costs.
Palomar Regulated Bias Schematic Elite 450HD
- Thread starter 443
- Start date
Yes in the Palomar amp using this Bias circuit the 7805 had a little TO-220 heatsink attached to it. Having dealt with these 7805, 7812, 7815 type regulators ALOT in my past lines of work, they take the dropping voltage really well, even without a sink, even in fully encapsulated package form TO-220F. I would think in any application though, it would be better for sure, to make sure it's heatsinked, just for accuracy and longevity sake.
You may want to "stack"...
A particular reason for that is for RF/EMI wandering and Resonance - there are stories of how people wanted to consolidate filtering into one giant cap value - ok to a point but you lose a reactive element that you can play with in your favor when it comes to filtering.
You already know large value farads take time to release as was as replace a charge and on top of that to stay stable.
Unless you're using a bucking converter or are into switching - the "line for line" cap filter could be consolidated, but instead of one large value - I'd try two smaller values based on a ratio of 1:4 - Say you want to run a 2.2uF - that's fine but are you safe? Let's go with something to keep the line powered "properly" so you don't lose a chunk of amps just trying to fill up the bank of caps...
Just a thought or two - I mean nothing serious but if you have noise issues or using heavy modulation somewhere - the power supply will need some "Breakpoints" to help remove the noisy sections it may have.
Some simple stuff, because when you have that many watts - it's not going to be easy to simply place one "cap" in, and be done with it - you'll need to scatter these around as you need to. To help keep the buss lines quiet and since you're looking into using this for Bias - a good practice is to filter the line well - even after the branching to prevent feedback paths from ruining your work.
A particular reason for that is for RF/EMI wandering and Resonance - there are stories of how people wanted to consolidate filtering into one giant cap value - ok to a point but you lose a reactive element that you can play with in your favor when it comes to filtering.
You already know large value farads take time to release as was as replace a charge and on top of that to stay stable.
Unless you're using a bucking converter or are into switching - the "line for line" cap filter could be consolidated, but instead of one large value - I'd try two smaller values based on a ratio of 1:4 - Say you want to run a 2.2uF - that's fine but are you safe? Let's go with something to keep the line powered "properly" so you don't lose a chunk of amps just trying to fill up the bank of caps...
Just a thought or two - I mean nothing serious but if you have noise issues or using heavy modulation somewhere - the power supply will need some "Breakpoints" to help remove the noisy sections it may have.
- A secondary benefit to this is to make the Regulator more of a "Regulator" not a noise pit trying to suck all the noise on the line unto itself and provide pure "unfettered" DC as the same time.
Some simple stuff, because when you have that many watts - it's not going to be easy to simply place one "cap" in, and be done with it - you'll need to scatter these around as you need to. To help keep the buss lines quiet and since you're looking into using this for Bias - a good practice is to filter the line well - even after the branching to prevent feedback paths from ruining your work.
Last edited:
The thing to remember about the 78xx regulators is that they contain a lot of amplifier gain. An error amplifier, yes. But gain. They will oscillate if the output pin doesn't have a sufficiently low AC impedance. Years ago we found a 10uf cap the least-troublesome default.
These regulator parts have been made by a great many vendors now over the years. Subtle differences in behavior from one source to the next seem likely. If you made an exhaustive study of parts made by 20 different sources, you might find the "sweet spot" for that capacitor value is not the same for all of them.
I do know that it needs to be very, very close to the chip's output and common pins to be effective. Never had a stability issue with 10uf.
Knock on wood...
73
These regulator parts have been made by a great many vendors now over the years. Subtle differences in behavior from one source to the next seem likely. If you made an exhaustive study of parts made by 20 different sources, you might find the "sweet spot" for that capacitor value is not the same for all of them.
I do know that it needs to be very, very close to the chip's output and common pins to be effective. Never had a stability issue with 10uf.
Knock on wood...
73
443 the diodes in the palomar bias circuit are not touching anything to make them track temperature because palomar amps are put together by folk who have ZERO clue about electronics,
the circuit as implemented does the opposite to whats needed as temperature rises,
its as if a computer designed the circuit but clowns ass embled them,
unsoldering the diodes & putting them under the board against the heatsink with a blob of thermal paste gives the circuit some thermal tracking,
That circuit won't be as stiff regulated as the motorola circuit but it will work ok to remove audible switching distortion on ssb,
the tip31 conducts harder as it heats up increasing idle current, two diodes in series seem to compensate for the outputs and the tip31 pass transistor better than a single diode in that style circuit,
BUT they can't track temperature if they don't touch the heat source, either the pill caps or heatsink close as possible to the pills.
crusher good to see you still out here
the circuit as implemented does the opposite to whats needed as temperature rises,
its as if a computer designed the circuit but clowns ass embled them,
unsoldering the diodes & putting them under the board against the heatsink with a blob of thermal paste gives the circuit some thermal tracking,
That circuit won't be as stiff regulated as the motorola circuit but it will work ok to remove audible switching distortion on ssb,
the tip31 conducts harder as it heats up increasing idle current, two diodes in series seem to compensate for the outputs and the tip31 pass transistor better than a single diode in that style circuit,
BUT they can't track temperature if they don't touch the heat source, either the pill caps or heatsink close as possible to the pills.
crusher good to see you still out here
Thanks for clearing that up Bob, I figured as much. Shame they make stuff, but they don't do it "right". I had no hand in that, only made the schematic from looking at the board/circuit to share it and get opinions on it.
One question, would the small 1N4148 diodes be any better to use, or should stick with the 1N400X series diodes? One is glass, the other is plastic case. Which would be better to use for right at the transistors/mounting arrangement?
I’ve have good luck using 1N4007 diode. I have used a file and file down 1 side of diode just to give it a little more area that touches heatsink. Tracks temperature a little better. I then use a small piece of strap or brass and solder across/ over diode to clamp down onto surface.
Like Crusher I also file a flat on the diodes to get better thermal contact, you just have to make sure you get them sat flat against the heatsink.
4000 series work fine, some hf sets use plastic & some like older kenwoods use smaller glass diodes in a similar cirtcuit,
either type work, i have not tested lots of different diodes to see which if any keep idle current more stable,
you can also use a transistor in place of the diode as seen in the motorola 723 circuit.
4000 series work fine, some hf sets use plastic & some like older kenwoods use smaller glass diodes in a similar cirtcuit,
either type work, i have not tested lots of different diodes to see which if any keep idle current more stable,
you can also use a transistor in place of the diode as seen in the motorola 723 circuit.
Hi All,
I'm still collecting parts to rebuild 900HD (above). I decided to use MRF422 as I already have a PSU (28v), they're cheaper and less likely to counterfeit. I'm waiting on delivery for last pair now. Got first pair <$40. Should be a drop-in replacement, right?
That brings me to Bias, a must for SSB.
The original 7805/Tip31 worked well in the past but I'm tinkering with the Bias circuit. I replaced the 7805 regulator with a LD1117 {3.3v} to put me closer to .6v. Motorola AN762 speculates about replacing the 2 diodes in-series for thermal tracking ("TT") by 'replacing the diodes with the base-emitter junction of a 2N5190 laying flat near RF Q’s to track temp changes.' See: http://www.communication-concepts.com/content/AN762/AN762_Application_Note.pdf
Beats heck out of filing the side of a silicon diode flat which was the competing method of adding TT. I'll keep this thread posted of my progress.
Ned, NYC
I'm still collecting parts to rebuild 900HD (above). I decided to use MRF422 as I already have a PSU (28v), they're cheaper and less likely to counterfeit. I'm waiting on delivery for last pair now. Got first pair <$40. Should be a drop-in replacement, right?
That brings me to Bias, a must for SSB.
The original 7805/Tip31 worked well in the past but I'm tinkering with the Bias circuit. I replaced the 7805 regulator with a LD1117 {3.3v} to put me closer to .6v. Motorola AN762 speculates about replacing the 2 diodes in-series for thermal tracking ("TT") by 'replacing the diodes with the base-emitter junction of a 2N5190 laying flat near RF Q’s to track temp changes.' See: http://www.communication-concepts.com/content/AN762/AN762_Application_Note.pdf
Beats heck out of filing the side of a silicon diode flat which was the competing method of adding TT. I'll keep this thread posted of my progress.
Ned, NYC
Using a transistor's base-to-emitter junction to track temperature tends to be more accurate than the diode method. More complex and expensive, but it does perform better, especially at high temperatures.
73
73