No, 0.5V (500mV) is -6dBV. 1V is 0dBV by definition.
Many regulators have a minimum load requirement. Many others work even better with a relatively heavy load. So no signs of the apocalypse in your observations! The 0.2V difference in the rails will make no difference to your circuit- the references are all to ground, and a small asymmetry in the rails doesn't upset that.
I don't think anyone has touched base on this but you don't have to have a 48 volt supply for Phantom power. Most mics will work on far far less than that. I have several Phantom power boxes that are +9 or +18 volt. The industry norm is a 48 volt supply but keep in mind that some mics will power up and work on as little as a 1.5 volt battery and some mics require a +90 volt supply. If I'm not mistaken either a + supply or a minus supply will work
I have played with both. The idea is to get a same supply down both the pin 2 and pin 3 of the mic line to the mic. Its amazing what I run into in the commercial audio field on a day to day basis.
I have played with both. The idea is to get a same supply down both the pin 2 and pin 3 of the mic line to the mic. Its amazing what I run into in the commercial audio field on a day to day basis.
Thanks Sy.
I did some reading and have seen suggestions that the LM317 needs 10mA or more between Vout and Adj. Since Vref is 1.2-1.3V, one needs ~120 ohms or less for the out to adj resistor to achieve this. With a 240R as shown on the Datasheet sample circuits, one has too low of an adj current. I'll play with that tonight and see if it changes anything
My mouser order should be waiting for me when I get home tonight some hopefully I'll make some progress soon.
6BG6BA, thanks for that info. I had considered using a lower phantom voltage but read In a few places that many newer and "cheaper" mics did not perform as well. It only takes two extra parts for ms to go to 48 volts instead of 24-35 so it's no biggie price or space wise.
Thanks for the info!
I did some reading and have seen suggestions that the LM317 needs 10mA or more between Vout and Adj. Since Vref is 1.2-1.3V, one needs ~120 ohms or less for the out to adj resistor to achieve this. With a 240R as shown on the Datasheet sample circuits, one has too low of an adj current. I'll play with that tonight and see if it changes anything
My mouser order should be waiting for me when I get home tonight some hopefully I'll make some progress soon.
6BG6BA, thanks for that info. I had considered using a lower phantom voltage but read In a few places that many newer and "cheaper" mics did not perform as well. It only takes two extra parts for ms to go to 48 volts instead of 24-35 so it's no biggie price or space wise.
Thanks for the info!
Dropping the phantom voltage reduces sensitivity and hence S/N. The loss will depend on the particular mike, and may or may not be noticeable, but TANSTAAFL applies.
I tried switching the ADJ to Vout resistor to 100R, which should give a current of 12-13mA. It did not affect the final unload voltage with the SMPS wall warts for a source. Will a torroidal transformer have the same issue?
If a circuit that does this is connected to a load drawing at least 40mA, would the voltage be high when it is first turned on or is a load instantly seen? My concern is that the unloaded voltage is higher than the max voltage for the preamp IC.
Thanks!
If a circuit that does this is connected to a load drawing at least 40mA, would the voltage be high when it is first turned on or is a load instantly seen? My concern is that the unloaded voltage is higher than the max voltage for the preamp IC.
Thanks!
The reg will limit the voltage. FWIW, I use solid state regs for my tube amps and preamps- with protection diodes, I've never lost one in use (test bench accidents are another story).
I see my mistake now.. I was originally using the wrong equation to find the voltage based on the Adj and Vout to Adj resistors. Somewhere along the line forgot that it is 1.25(1+R2/R1). I was using R2/R1 and not getting why my voltage was so higher than expected. If I do the math right I get what I expect regardless of load.
Below is the final board layout. Sorry there are no values on it. Note, the LM317/337 trimmer resistors are simply placed for potential need. I obviously don't need the top one and it won't be on the version I etch. The two middle ones will most likely be ignored and a fixed resistor used instead. The last set of holes on the dual supply is for a load resistor if needed. I only had room for two large caps and a singe resistor CRC, filter on the rectified but preregultor part of the dual supply. My plan is to use only the first cap and add later if necessary.The D's near the adj caps are protection diodes. The is on jumper from the positive rail adj pin down to the resistor/trimmer. That jumper is not marked. Looking at this you, you may ask why I did not go right to the ground in the phantom area. I have no idea since I draw a trimmer there. I assume that would be OK?
Thanks all and I'd love critique of the board as I'm sure it could be much better.
Below is the final board layout. Sorry there are no values on it. Note, the LM317/337 trimmer resistors are simply placed for potential need. I obviously don't need the top one and it won't be on the version I etch. The two middle ones will most likely be ignored and a fixed resistor used instead. The last set of holes on the dual supply is for a load resistor if needed. I only had room for two large caps and a singe resistor CRC, filter on the rectified but preregultor part of the dual supply. My plan is to use only the first cap and add later if necessary.The D's near the adj caps are protection diodes. The is on jumper from the positive rail adj pin down to the resistor/trimmer. That jumper is not marked. Looking at this you, you may ask why I did not go right to the ground in the phantom area. I have no idea since I draw a trimmer there. I assume that would be OK?
Thanks all and I'd love critique of the board as I'm sure it could be much better.
I sure did, thanks Tony!
BTW, what size resistors are you using to set the voltage and do you have a load resistors after the regulator?
MicPrePSUfinal by Dennis Dietz, on Flickr
BTW, what size resistors are you using to set the voltage and do you have a load resistors after the regulator?
MicPrePSUfinal by Dennis Dietz, on Flickr
My voltage set resistors are 10K and a 50K pot, but that is because I'm using the Fred Dieckmann mod (with a bc560c from adj to ground, base is fed from the junction of the 10K and 50K pot. Effectively what this does (as far as I can tell) is makes the adjust pin capacitor much more effective as the 10K resistance (compared to the normal 240 ohms) means that the RC filter created has a MUCH lower corner frequency.
There are a few disadvantages to this method however. One is that the formula for calculating the resistor values no longer works (and It doesn't appear to be linear so I haven't worked a new one out). and also as someone else mentioned there is a thermal drift issue, which isn't a real problem if your voltage isn't super critical, but I have noticed the voltage varies (we are talking about maybe a few tenths of a volt) as the circuit warms up.
I was planning on having about 250 ohms load resistors (which for the 10V I'm going to be running is about 40mA) my actual circuit I'm estimating will draw around 160mA so that totals around 200mA, after reading SY's comment about the higher current, and also a link that someone else provided (might have been in my measurements thread) I'm wondering if I should drop that resistance a bit... I just bought some 523 ohm resistors (was what was available, the value I wanted had an August lead time!) and was planning on using two in parallel to handle the power dissipation. but I may end up running three or four in parallel.
I'm having some difficulty reading your layout, but the tips that I have come from the datasheet.
That is: make sure your upper resistor in the divider is connected as close to the reg output as possible. The datasheet actually suggests connecting it directly to the device.
The 0V side of the divider should be as close to the load as possible as well (from the datasheet)
The other thing (also based on the datasheet) I would suggest would be to have the regulators as close to the output as possible.
However I really have no idea on layout, and as yet haven't received any feedback on my proposed layout, so take everything I say with a large grain of salt 😉 However you might want to read this post to see where I'm coming from http://www.diyaudio.com/forums/power-supplies/184068-psu-rc-multistage-filtering-2.html#post2487825
Tony.
There are a few disadvantages to this method however. One is that the formula for calculating the resistor values no longer works (and It doesn't appear to be linear so I haven't worked a new one out). and also as someone else mentioned there is a thermal drift issue, which isn't a real problem if your voltage isn't super critical, but I have noticed the voltage varies (we are talking about maybe a few tenths of a volt) as the circuit warms up.
I was planning on having about 250 ohms load resistors (which for the 10V I'm going to be running is about 40mA) my actual circuit I'm estimating will draw around 160mA so that totals around 200mA, after reading SY's comment about the higher current, and also a link that someone else provided (might have been in my measurements thread) I'm wondering if I should drop that resistance a bit... I just bought some 523 ohm resistors (was what was available, the value I wanted had an August lead time!) and was planning on using two in parallel to handle the power dissipation. but I may end up running three or four in parallel.
I'm having some difficulty reading your layout, but the tips that I have come from the datasheet.
That is: make sure your upper resistor in the divider is connected as close to the reg output as possible. The datasheet actually suggests connecting it directly to the device.
The 0V side of the divider should be as close to the load as possible as well (from the datasheet)
The other thing (also based on the datasheet) I would suggest would be to have the regulators as close to the output as possible.
However I really have no idea on layout, and as yet haven't received any feedback on my proposed layout, so take everything I say with a large grain of salt 😉 However you might want to read this post to see where I'm coming from http://www.diyaudio.com/forums/power-supplies/184068-psu-rc-multistage-filtering-2.html#post2487825
Tony.
Last edited:
Thanks for th reply Tony. Sorry the layout is hard to read? May I ask what was hard; color, lack of values, etc?
Thanks also for the tips. I think I've done all of those, but only but luck🙂. I hav always had concern that my board layouts are not the best. I tend to layout in a linear fashion, like a schematic and I think my creativity suffers for lack of experience. My biggest concerns are with grounding, where and how I run the ground traces. Beign limited to generally single sided compounds that since I make my own boards at home. I've been successful with double sided at home but the lack of plating the through holes makes for lots of jumpers everywhere.
My two biggest concerns at this point are the current set by the adj to Vout resistor (including the overall load of the circuit) and layout of the board in terms of grounding traces.
Regarding the necessary current of the circuit, I've not seen anything on the datasheet except that they always mention 10mA in the performance specs, but all the schematics on the sheet use 240R for the resistor, which leaves a current of 5-5.4mA. Seems weird to me. They never mention or show load resistors nor have I seen current recommendations. Sy has mentioned that higher current is better, I'm of course inclined to believe and trust him.
The practical answer for my needs here is how much of this really matters🙂. Sy has suggested that the noise values you've found and posted here would be inaudible in a recording situation. I suspect I'll find out very soon🙂
Thanks also for the tips. I think I've done all of those, but only but luck🙂. I hav always had concern that my board layouts are not the best. I tend to layout in a linear fashion, like a schematic and I think my creativity suffers for lack of experience. My biggest concerns are with grounding, where and how I run the ground traces. Beign limited to generally single sided compounds that since I make my own boards at home. I've been successful with double sided at home but the lack of plating the through holes makes for lots of jumpers everywhere.
My two biggest concerns at this point are the current set by the adj to Vout resistor (including the overall load of the circuit) and layout of the board in terms of grounding traces.
Regarding the necessary current of the circuit, I've not seen anything on the datasheet except that they always mention 10mA in the performance specs, but all the schematics on the sheet use 240R for the resistor, which leaves a current of 5-5.4mA. Seems weird to me. They never mention or show load resistors nor have I seen current recommendations. Sy has mentioned that higher current is better, I'm of course inclined to believe and trust him.
The practical answer for my needs here is how much of this really matters🙂. Sy has suggested that the noise values you've found and posted here would be inaudible in a recording situation. I suspect I'll find out very soon🙂
At 5mA, most 317s will function. At 10mA, all 317s will function, so I usually use 120R for that resistor. I found this out the hard way. I also found out the hard way that not all datasheet circuits work correctly- one notorious example is the Maida appnote for using the 317 in high voltage circuits (hint: what's the drop across the input resistor at max rated current?).
You don't have to trust me about the desirability of sinking a lot of unused current (100mA or more)- Erroll Dietz's data speaks for itself and that's where I learned about it.
Tony's suggestions about layout are spot on- keep the 120R (if that's what you use) tight to the 317, return the output cap (e.g., the - lead for the positive rails) to as close to the load as possible.
You don't have to trust me about the desirability of sinking a lot of unused current (100mA or more)- Erroll Dietz's data speaks for itself and that's where I learned about it.
Tony's suggestions about layout are spot on- keep the 120R (if that's what you use) tight to the 317, return the output cap (e.g., the - lead for the positive rails) to as close to the load as possible.
Thanks Sy. I appreciate it. I should say I don't necessary blindly follow anyone, but when i say I trust you, I mean I trust your experiences and your ability to interpret date ( like the Erroll Dietz) better than me🙂.
As for the tips, I've tried to keep the parts as close as possible. My resistors you mention are as close to the output pin ofthe 317 as the body of each will allow. While that was purely luck, I do generally try to keep everything as close as possible, with in reason of course.
May I ask what you think of my board layout, at least if there any potential major problems?
As for the tips, I've tried to keep the parts as close as possible. My resistors you mention are as close to the output pin ofthe 317 as the body of each will allow. While that was purely luck, I do generally try to keep everything as close as possible, with in reason of course.
May I ask what you think of my board layout, at least if there any potential major problems?
I have an innate inability to read and interpret circuit boards. That's why I do point to point. 😀 The set resistor's proximity is much less important than the feedback resistor- if they're both close, fine. The return end of the current setting resistor should be close to the load, if possible.
Thanks Sy. I have a much harder time with point to point, everything always looks the same to me. Reading a pcb is like reading a map, for me, and I think I helps me that there are actual "road" with "landmarks" I can remember.
Sorry, but when you say current setting resistor, are you referring to the lm317 resistor from Vout to Adj, or a load resistor from either rail to ground?
Sorry, but when you say current setting resistor, are you referring to the lm317 resistor from Vout to Adj, or a load resistor from either rail to ground?
Last edited:
My mistake, sorry- I meant "voltage setting resistor," i.e., the resistor from the adjust pin to ground.
Hi Dennis
I think the problem is a lack of contrast, a darker colour for the component overlay and writing would probably make a big difference 🙂
Tony.
I think the problem is a lack of contrast, a darker colour for the component overlay and writing would probably make a big difference 🙂
Once again not my strong point! I only learn from experience, and this is an area where I am lacking in that regard.
Trust a Texan? 😀 One thing I don't know if you have considered, you can just put a resistor (of ample wattage) from the output to 0V to get your minimum load current, it doesn't have to be through the adj resistor 🙂 That is what I will be doing. As for placement of this resistor I haven't 100% decided yet...
Most likely! my measurements were with 10X gain which means that they were around -120db to start with.... I'm just one for trying to wring every last drop out of something because I can, and just in case it might make a difference, even if most people will say it wont 😉 Even if I can measure a difference, I have strong doubts that I could hear it.
Tony.
Last edited:
- Status
- Not open for further replies.