Help/advice/guidance translating schematics onto a pcb

[Brian Roth]

It is a bit odd that the VPR Alliance spec says 5 mA max for the 48V rail while the "official" IEC standard mentions up to 7 mA (10mA max) for any given mic or other device. The VPA spec ignores the existance of "hungry" mics. However, most mics draw 5 mA or less.

Keep in mind we have two "branches" for the 48 V rail after it passes through the phantom switch.

The primary path is through the pair of 6k8 resistors which provide power to the mic. The absolute worst case current would be approx 14 mA with a fault condition where XLR pins 1, 2 and 3 were all shorted together (totally crapped out mic cable!).

[Sidebar: The 6k8 resistors will dissipate around 0.3 Watts under the bad fault condition so need to be rated at least 0.5 Watts. I've seen more than a few consoles that had been in service for many years which had random 6k8/0.25 Watt resistors "browned" due the presence of a bad cable at some point.]

The other 48V path in the proposed design (and the CAPI schemo) only makes the lamp illuminate. In this case (10K in series with the LED which will have approx. 2 Volt drop across it), approx. 4.6 mA will flow with the lamp "on". If no phantom flows to a mic, the module will draw 4.6 mA. A hungry mic drawing 10 mA means the total module draw will be 14.6-ish mA. A boogered mic cable (worst case condition) raises the total to approx 18.6 mA. Note that the LED will still light even with the boogered cable.

Rambling answer to @djmiggymigz IMHO, the 4.6 mA draw in the CAPI design may be too high to account for a wimpy 48V power supply and whatever. As has been mentioned in this thread, a high efficiency LED (and a higher series resistance) should be a good solution to CYA.

Bri

PS, for jollies I looked in the JKL catalog for 48V incandescent bulbs. Found one that draws 53 mA! lol

 

[Cqwet Dbdfte]

Don't be afraid of using 6.8k wire-wound resistors, budget and space permitting. 3W are 1/2" long.
Not as good as a choke, but trying to;-)

 

[Cqwet Dbdfte]

Adding a green LED in series with supply would indicate that the load is applied. If that is of any value.
A weak current from LED cathode to ground could give some indication of power applied.
A typical LED will drop out about 1.8V some types over 3V.
Not sure if the voltage loss is tolerable.
The intensity of the LED would indicate load current, opens and shorts.

 

[Brian Roth]

Adding a green LED in series with supply would indicate that the load is applied. If that is of any value.
A weak current from LED cathode to ground could give some indication of power applied.
A typical LED will drop out about 1.8V some types over 3V.
Not sure if the voltage loss is tolerable.
The intensity of the LED would indicate load current, opens and shorts.

Unless I'm missing something, the LED in series won't illuminate unless there is a load from the mic. Hence, it won't indicate the presence of 48V until a mic is plugged in, which seems to defeat the purpose of the indicator .

Bri

 

[Alva Goldbook]

I learned on Eagle, but moved over to KiCad. There's a few easier programs you can use for simple stuff like this. Learning how to generate gerbers on something like KiCad can be quite the undertaking, it's not something you learn over night, but if you dig electronics, you might as well start now. Watch some how tos on Youtube and just go for it. Send your gerber off to a board house and see how it turns out. The best way to learn is by doing. And expect it not to go right the first time. Happens to all of us. That's why we prototype things. I can do up a small order of boards for something simple for like 20 bucks. So if it doesn't work, no biggie, fix your errors and try again. The important point is to take the plunge!

 

[Cqwet Dbdfte]

Unless I'm missing something, the LED in series won't illuminate unless there is a load from the mic. Hence, it won't indicate the presence of 48V until a mic is plugged in, which seems to defeat the purpose of the indicator .

Bri

You missed my remark of "a weak current from the cathode".
Add a high value resistor to to ground, indicate 48V on, then plugging in a load would show a different intensity.
Cheers

 

[djmiggymigz]

a high efficiency LED

Okay so when I'm looking at the specs/datasheets for LED's in regards to getting a "high efficiency" LED, I should be looking at the Forward Current correct? I.e if an LED states a forward current of 20mA, that's more or less the current that it will be drawing?

Thus, looking for a high efficiency one would look something like a Forward Current of around 5-10mA?

 

[Cqwet Dbdfte]

Okay so when I'm looking at the specs/datasheets for LED's in regards to getting a "high efficiency" LED, I should be looking at the Forward Current correct? I.e if an LED states a forward current of 20mA, that's more or less the current that it will be drawing?

Thus, looking for a high efficiency one would look something like a Forward Current of around 5-10mA?

Stated current is the value corresponding to stated light output.
The LED will draw as much current as you let it. Or until it burns up, if too much.
Efficiency is typically stated as mcd, millicandela, when associated with a mA number would roughly give the efficiency.
WP710A10LZGCK is rated 1100 mcd and 2.65V, @ 2mA.

 

[djmiggymigz]

Efficiency is typically stated as mcd

Ohhh okay I completely glanced over that number/spec,, didn't know what it meant! So is efficiency determined by the combo of mcd, Forward current, as well as forward current?

Specific example: 20 mcd, 1.9V, 20mA = ??? I think 20mA is a lot for my specific case as well as maybe the 1.9V.

The LED will draw as much current as you let it

Is this determined by the series resistor prior to the LED??

 

[Cqwet Dbdfte]

Yes, LEDs are current driven devices, forward voltage does not change much.
A resistor will easily set the current.

 

[djmiggymigz]

Okay back to the newbie questions that I'll look into on my own but also going to ask here!

Does the resistor -> current relationship work something like this: The series R only lets a specific amount of current/V thru it. Meaning depending on the R value, the LED is only getting a calculated portion of the current being fed to the resistor. Then, since it has to go somewhere, the rest of the current is fed back down the main path towards the mic pins?

Essentially, depending on the LED I get, I can alter the series resistor value to let through as little current as possible (to the LED), like just enough for the light to light up well. And thus, minimize the amount of current being taken away from the actual Phantom power being fed to the mic?

 

[Brian Roth]

@djmiggymigz In this circuit, the amount of current flowing through the LED is controlled by the resistance in series with the LED and is calculated using Ohm's Law. As I mentioned in post #61 above, the +48V coming in via pin 15 splits into two paths after the phantom switch. One path feeds the resistor+LED, the other path feeds the 6k8 resistors for the mic powering. Neither path "knows" what the other path is pulling in terms of current. However, the sum of both currents is drawn via pin 15. It's not like the LED "steals" current going into the mic powering path....it just is added into the total that is drawn through pin 15. The 48V power source feeding pin 15 must be capable of supplying the total of LED current plus mic current. The VPA standard of 5 mA or the more reasonable 7/10 mA IEC standard (per slot) can be thought as a "budget" for both requirements.

So, the idea here is to reduce the the amount of current pulled through the resistor+LED path to minimize how much of the "budget" is used up. And yet, you want the LED to be visible. This could be a whole different thread! lol

As technology has progressed, LEDs can emit a LOT more light in year 2024 vs. 1984 for a given amount of current flow. That's good for our situation! Some colors are more visible to the eye. LEDs have a "viewing angle" spec....do you have to be looking directly down the center line of the LED to see it, or can you be off-axis and still see the indication.

And, how the LED is attached to the control panel will have an effect. If the LED is recessed in a hole through the panel, it will be harder to see. There are inexpensive plastic "gizmos" that hold the LED in place and also disperse the light. Example:

https://vcclite.com/product-category/panellensesclips/

Finally, in recent years I've been seeing LED power indicators on equipment which have an insanely bright blue light! But, I have no idea how much current they draw.

Bri

 

[Cqwet Dbdfte]

The LED I suggested, is at the top end in efficiency, 100uA may be enough .
(48V - 2.75V) / 0.0001A = 450kOhm or so may be bright enough.
3uA it starts lighting up.
You don't want an indicator light make you go looking for sunglasses.
The development of white LEDs (blue!) has led to ever increasing efficiencies in LED tech.

 

[Brian Roth]

The LED I suggested, is at the top end in efficiency, 100uA may be enough .
(48V - 2.75V) / 0.0001A = 450kOhm or so may be bright enough.
3uA it starts lighting up.
You don't want an indicator light make you go looking for sunglasses.
The development of white LEDs (blue!) has led to ever increasing efficiencies in LED tech.

Yeah, this evening I was playing "what if" with the Mouser LED selector tool:

https://www.mouser.com/c/optoelectronics/led-lighting/led-emitters/standard-leds-through-hole/

I was using the 2mA choice and then looking at mcd, viewing angle and color. There are more than a few combinations available that should work well.

Bri

 

[Cqwet Dbdfte]

Found a 3000mcd 2mA, WP7113LZGCK
5mm, Too bright?

 

[Brian Roth]

Indeed....KIngbright is the Jolly GREEN Giant of LEDs! <g>

I notice their datasheets always mention concern about static discharge....perhaps a VCC "lens" as I linked above would be wise????? <shrug>

Bri

 

[Cqwet Dbdfte]

All blue and some green LED have the same warning, once installed should be OK.

 

[MidnightArrakis]

Found a 3000mcd 2mA, WP7113LZGCK
5mm, Too bright?

[Too bright?] -- !!!

/

 

[djmiggymigz]

Hello again everyone!!

Been working on this almost constantly since my last reply! Here is my update and where the project stands at the moment!

Got the pcb's from JLC, VU from Aliexpress, and a blank L-bracket and faceplate from CAPI. Got everything installed and finally got brave enough to plug it in last night aaand it works!! Mostly! Input gain is good, pad is good. However, my polarity wasn't working nor was the low cut! Also my output attenuator is working but opposite than I intended?? Currently clockwise rotation attenuates which my brain absolutely hates but it's working so still a win!

It was so hard to wrap my head around how to layout and drill out the L bracket and faceplate! I didn't pay much attention to the physical laayout dimensions when designing the pcb so that took me a good few days. This iteration is mucked up but I saw it through for the practice. I have a new bracket and faceplate en route and thanks to my jeweler roommate i now have the tools to cut and polish up the holes real nice.

Sadly though, I think just one more PCB iteration is needed and that should do it. I'm going to scrap the Lowcut filter since it didn't work and the button is kind of taking up too much space. I think it'll be a bit better overall to scrap it and free up some more real estate. That'll also allow me to move the 2 remaining buttons down (they were rubbing against the input knob) and move the input down (It was too close to the VU meter). Also found that I need to space my resistors out a bit more for it to sit flush and look nice (the vishay resistors I used are thicker than most). Oh and I did some looking at other polarity schematics and I think I just had it wrong in this iteration so I need to fix that as well.

The VU also wasn't working. It was lighting up just fine but wasn't responding to signal at all. Did some fiddling and the TL072 IC started to smoke lmao so that'll be a little something to figure out!

Overall though, I'm super excited. The fact that it passed signal was so gratifying and seeing it physically come together felt like I had won a boss fight in a video game! I just need to iron out some wrinkles and it'll be ready for a life in the studio. Huge thanks to everyone here for helping me through this!! Hope y'all get some joy out of seeing it come together because I'm ecstatic!!

Also for anyone that's curious about the cost, this comes together for right around $200 top to bottom! Not bad considering the current API iteration is around $800 new.

 

[djmiggymigz]

Oh also, re: the whole conversation regarding the phantom power circuit and LED/current draw etc. I went with the same idea that the CAPI units use and it works perfectly with my Fredenstein Bento chassis! Do I understand it? No definitely not. But it works! I'm sure I'll come to understand sometime this year hahah