John Curl's Blowtorch preamplifier part II

Status
Not open for further replies.
Well back to audio.

Attached is my 50 ohm load from a CT 24V transformer. I used four 4700 uF capacitors and three 1 ohm resistors. It went from all the capacitors in parallel to a four stage CRCRCRC filter. As a CRCRC filter it did not seem to matter where the 9400 uF was positioned. Of course the four stage filter was the best. Total voltage loss was 6%. So the conclusion seems to be calculated what the worst case into the filter will be and the minimum you need out of it and use a three or four stage filter that uses up the voltage.

ES

Ed, what are you looking at? Filtered out dc with one diode? What schematic?

jn
 
didn't notice

If there's no sensitory difference, summer or winter, means 100% success score of the repair job.
Nerves are not single solid core, but strands. Bigger the function, the more strands.
The lower the function, fewer strands, higher chance of conductivity issues, and/or under/over-sensitivity.
Double or nothing is my middle name.

(I read latin, but from a different bible)
 
If there's no sensitory difference, summer or winter, means 100% success score of the repair job.
Nerves are not single solid core, but strands. Bigger the function, the more strands.
The lower the function, fewer strands, higher chance of conductivity issues, and/or under/over-sensitivity.
Double or nothing is my middle name.

(I read latin, but from a different bible)

Ah. So if there's high conductivity, then maybe I'm suffering skin effect???

Seriously, thanks for the info..really appreciate it.


2 Diodes, CT grounded. Capacitor and series resistors are the variable. Output is across the 50 ohm load resistor. 0dB is 1 VRMS.

Did you match the diodes? There's an awful lot of 60 and odds there..

What speed are the diodes?

jn
 
Clock Makers Tool Kit ?...

henry-studley-tool-chest.jpg

Henry Studley's Tool Chest. - Neatorama
Dan.
 
Member
Joined 2004
Paid Member
jn:

I thought you were building one of these. Its on a par with what you do for your day job. Its still running which is a good indication of the life expectancy of wooden gears. He did not have access to a multi-axis CNC machine, or even a router. Just lots of time and patience.
 

Attachments

  • harrison mov front plate removed.jpg
    harrison mov front plate removed.jpg
    281 KB · Views: 188
Well, I don't make wooden clocks or use milling machines, but it seems kind of 'decadent' doesn't it? I mean, do they keep better time? Of course not, but I will not criticize your wanting to put YOUR time, effort, and money into it.
I personally like to invest in hi end audio, instead, but not as much as I used to. Of course, the stuff I design, I can get cheaply or for free, but the other stuff, like cables, speakers, turntables, tuners, etc, I have to 'scrounge' for, usually buying them used, and a bit out of date. I know the difference, but I just cannot justify it.
For example, I have a Linn turntable that I got in trade for a Vendetta upgrade. It is not on the 'recommended' list in Stereophile this month, or for several years, but it does OK. Instead, its updated equivalent for $5,000 is given a high rating. I just cannot afford the updates.
 
The pictures shown are with something fancy, (http://www.mouser.com/ds/2/389/CD00050205-112771.pdf) pretty much the same with GI 1N4004 and a few others. The 60 hz is from the transformer unbalance. It is a Signal flat pack, the worst of the bunch.
Ah, ok. So your diodes are not vf matched. That would also explain some of the 60 and odds. Try matching two diodes at op current dc to within a millivolt, see if that helps. Don't use pulse matching, as the match accuracy will depend on themal effects such as die attach integrity (voids in the braze between the aluminum and the moly for 4001's, and the hitemp solder between the schottky and the header).

If you have the same number of turns and still xfmr imbalance, perhaps working the lead dress may help. Replace the diodes with a matched set of resistors, no filter but 1k or so load, play with the leads until signal minimum.



Very nice, color me green. What's interesting is that Harrison was also a carpenter, he built wooden clocks. H1 apparently has wooden gears even though he used lignum v rollers on the pinions for reduced friction. Anyone at the observatory, take the tour of the clocks. Then gentleman who gave the talk when I was there was so awesome, and really knew his stuff.

jn:

I thought you were building one of these. Its on a par with what you do for your day job. Its still running which is a good indication of the life expectancy of wooden gears. He did not have access to a multi-axis CNC machine, or even a router. Just lots of time and patience.
Those are exactly the type of gears I'm fixing on occasion. If you look top center, the escape wheel is brass, the pallets are probably steel hardened at the tips. Lower left the count wheel is built onto the weight mandrel. Time side is the right side of the mechanism.

edit: ah, just noticed the link name is harrison. Cool. It looks like the teeth are actually dowels inserted in the gear wheel edges. They are too thin to survive if the grain cut across the teeth.

One issue I come up with on occasion is the teeth may not have been created using an indexing plate. When I put in a section of new teeth, I have to eyeball the new tooth profile...a 36 tooth gear's teeth may deviate from 10 degrees by almost a degree or two, requiring finess. Not having finess, I make do..:D

What amazed me the most is the level of technology achieved in the 1600's and 1700's, Invar was developed in 1750's I recall, as a consequence of temp dependence. I have a book called this modern clock, written 1905, it is still dead on the money.

jn
 
Last edited:
Well, I don't make wooden clocks or use milling machines, but it seems kind of 'decadent' doesn't it? I mean, do they keep better time? Of course not, but I will not criticize your wanting to put YOUR time, effort, and money into it.

First, they are a part of history. Man's use of available materials and thinking to measure time. The research in bettering time spun off huge benefits for man. Harrison's work saved countless lives at sea. I recommend the book Longitude.

Second, learning this stuff is an enrichment. Teaching a 12 year old how to fully disassemble and reassemble an 8 day mantel clock is in itself rewarding. While he may never do it again, he learned some new skills and confidence. We ALL should be instilling that in the next generations...

At the turn of the century, they were already making clocks which were 5 millisecond accurate per day, in an age when there was no such thing as a room thermostat...that is technology. (I've no idea how they proved that accuracy back then)

That is why I do not criticize you for wanting to put effort into audio. The goals, while different, are also the same.

jn
 
Ah, ok. So your diodes are not vf matched. That would also explain some of the 60 and odds. Try matching two diodes at op current dc to within a millivolt, see if that helps. Don't use pulse matching, as the match accuracy will depend on themal effects such as die attach integrity (voids in the braze between the aluminum and the moly for 4001's, and the hitemp solder between the schottky and the header).

If you have the same number of turns and still xfmr imbalance, perhaps working the lead dress may help. Replace the diodes with a matched set of resistors, no filter but 1k or so load, play with the leads until signal minimum.

jn

John,

I am using my worst transformer and diodes while exploring the filter network!

Even so the 60 hz is 108 db down from the DC output.

ES
 
John,

I am using my worst transformer and diodes while exploring the filter network!

Even so the 60 hz is 108 db down from the DC output.

ES

108 db down?? My wife can hear that from the kitchen doin the dishes..

So then it's filter issues only...ok..

For axial electrolytics, run a braid over the bodies of them and use the braid as the return current path. That will reduce mutual inductance, and make the capacitor as close to a 4 lead device as possible.

jn
 
Status
Not open for further replies.