Reflektor-D builds

Member
Joined 2018
Paid Member
Not been around here lately. Mainly because my Salas regs work nicely and I've learned a lot to not need to ask how to set them up.!
Grabbed an unused one out today as I'm building a new streamer. Thought I'd use it for an Allo Kali.
I had a few swerve balls and I did think maybe I'd be asking for help but managed to sort it. Hopefully some of this will impart some knowledge to any novices to help them fault find . Reading through these threads including any issues with builders projects is key to learning this game...IMHO!
So I set up for 5v with 2 green and a jumper. Kali tech manual says 5v 3A PSU. 5v I'm fine with but 3A?! In comparison Ian Canada Fifo is 200mA with supplied clocks...I'll aim for this.!
So initial test is suspect. 21R as test load. 3r3 as CCS setting R1. I get about 4v2 and no LEDs....hhmm. Something tells me from Ubib collapsing due to insufficient CCS to change R1. So I paralleled a big 21r from speaker XO, gives about 2r8. Now we have LEDs! But only 4v6 on the output. Tried a green LED in Rxjd. No change. Eventually saw that I could get 4v6 to 4v8 output with no LEDs in any position. Strange. Then deduced that the 21r load and asking for 5v was also requiring 240mA, which I wasn't providing with R1. So I replaced the test load with the Kali and all was well. 5v. Eventually removed the parallel 21r at R1 and still all good. Then asked the Kali to feed thru to power Ian Canada isolator underneath via the GPIO socket and all still good at 5.05v. Sinks on the Reflektor cool though so shall now up the CCS again.
Interesting afternoon!
IMG20230329171813.jpg IMG20230329171829.jpg
 
  • Like
Reactions: 1 user
diyAudio Chief Moderator
Joined 2002
Paid Member
Many times manuals just refer to SMPS wallwarts amperage reflecting the usual spec of commercially available models. Not to the actuall mA min/max to be consumed by the application.
Tech approach is to 5V set a bench power supply first and read the device's under test consumption behavior, then set Ref-D CCS above peak demand by its formula.
Setting the CCS level by trial and error R1 value is a hands on learning experience about actual applications demands. Current limiting live until sufficient won't hurt something as you saw. In effect its like coming up with a Variac.
 
  • Like
Reactions: 1 user
diyAudio Chief Moderator
Joined 2002
Paid Member
Bench PSUs offer a variable current limit too. For protection purposes. If you start very conservatively then increasing CC you simulate what you did this afternoon. A thing you don't get to know until moving to the dedicated PSU is the dissipation of your shunt reg build for how much extra mA allowance.
 
  • Like
Reactions: 1 user
diyAudio Chief Moderator
Joined 2002
Paid Member
You may find this general Reflektor negative topology of use as an example. Comes from my back then research & development archives. Keep in mind two positives arranged as pseudo symmetrical (stacked like batteries with virtual zero midpoint) are more alike than a mirrored components true ground symmetrical. Because NPN/PNP semis are not truly the same just inverted.

Reflektor negative sim test.jpg
 
diyAudio Chief Moderator
Joined 2002
Paid Member
@tamra

Hi, a busy build but it works alright obviously.
We generally prefer to at least have 5V DCrectified to DCout difference in the Ref-D and Ubib. For the CCS Mosfet to develop moderate internal capacitance and perform quick. But it still works at a lower difference. Tx secondary voltage also depends to a degree on local wall voltage average and in how much loss there is in the bridge diodes type too. How much DC difference there is now for you? The 1A Tx spec is fine for current in your application.