The matching transformer and toride T37-43 & DIRECT SAMPLING

Main forum to discuss RTL-SDR related topics.
Post Reply
P-40 Warhawk
Posts: 86
Joined: Thu Jun 26, 2014 11:55 pm

The matching transformer and toride T37-43 & DIRECT SAMPLING

Post by P-40 Warhawk » Tue Jul 01, 2014 12:00 pm

Hi to everyone .
I am new to this forum . I haven't found much talk about the T37-43 toroid and Independence on the secondary .
I have found a on line calculator for matching transformers . Since I am using 75 ohm cable , I used the numbers that It gave me for seventy five ohms on the primary . I am using the T37-43 toroid and I have done two dongles . One is 75 ohms to 200 ohms . 75 0hms to 500 ohms . I found that the higher the Independence on the secondary the higher the signal amplitude is , I get three s units more with 500 ohms Independence on the secondary than with 200 ohms Independence . I think that the RTL 2832 chip inputs are a very high Independence so from I have found you can use more than 200 ohms on the chips input . I would to know what other people have found ??????????????????????????
The on line calculator web site is . ... tching.php

P-40 Warhawk
Posts: 86
Joined: Thu Jun 26, 2014 11:55 pm

Re: The matching transformer and toride T37-43 & DIRECT SAMP

Post by P-40 Warhawk » Fri Jul 11, 2014 1:24 pm

Well I will try this again , I was hopping to find out if anybody was using any of the mini - circuits matching transformers for the direct sampling method on the rtl-sdr dongles . In the hobbyist's guide to rtl-sdr the author makes the statement that no one knows what the input Independence is for the rtl2832u chip
. Now I found a drawing on line that come from a person by the call sign YU3MA. In the drawing notes he states that a mini - circuits matching transformer ( t1-6t-kk81 ) works the best . I checked on EBAY and found the only mini - circuits matching transformer being sold was ( T36-1-kk81 ) , I wrote the seller and asked if he know what people were using it for . His answer was the he didn't know , but he was selling a lot of them . Dose anybody know anything about using the mini - circuits matching transformers on the rtl-sdr dongles ?????????????????

Site Admin
Posts: 2559
Joined: Mon Nov 19, 2012 11:54 pm

Re: The matching transformer and toride T37-43 & DIRECT SAMP

Post by rtlsdrblog » Sat Jul 12, 2014 5:49 am

I can't find the link again, but I remember there was a Japanese RTL2832 experimenter who also found that higher impedance's gave better results. He was using his own wound coils. As far as I know no one has really confirmed what the true input impedance is, and since the datasheet is not available we may never know for sure.

Posts: 2
Joined: Wed Jul 23, 2014 1:31 pm

Re: The matching transformer and toride T37-43 & DIRECT SAMP

Post by me1900 » Wed Jul 23, 2014 1:37 pm

I see 2 people have found it's a higher impedance. Enough for me. Maybe the dongle needs higher impedance feed lines for increased sensitivity. Ladder line might help. If someone could put a impedance tester of some kind to the antenna jack and match it. Maybe bridge the impedance of a 75 ohm coax feed line to match. If so, hey we've been doing it wrong all this time! ... -analyzers

Posts: 20
Joined: Fri Jan 19, 2018 10:37 pm
Location: Netherlands, Nuenen

Direct sampling dongle V3 modification; input impedance of RTL2832

Post by Brightnoise » Tue Jan 30, 2018 11:13 pm

Reading these posts I decided to remove the original transformer in my dongle (recent production; version V3) and to measure the input impedance of the baseband chip. Not with an impedance meter (overload!), but by adding and removing outside series and parallel impedances; complex and real. And this is what I measured: Ri differential is 6000 Ohms with a parallel capacitance of 12 pF. The 12 pF is not exact, since I measured it with an inductor with its own capacitance of a few pF.

Then I wanted to modify the lower frequency sensitivity loss. The 10 uH for antenna supply current, the diplexer component modification and the BGA2867 input and output modifications changed the response according expectations and did not give a new signal path problem. But then I found that the BGA2867 is used far out of the specified frequency range and easily can go into a DC latch-up situation. Far too easy! So I removed the BGA2876 also. Then I made a new transformer 77 uH primary and 288 + 288 uH secondary (for my ferrite transformer 6 on 12 + 12 windings on FairRite 2873002402). And now the frequency response is (1 MHz is reference 0 dB) -1 dB at 180 kHz, -3 dB at 95 kHz, -6 dB at 52 kHz, -10 dB at 32 kHz and -20 dB at around 12 kHz. No input amplifier in the dongle any more.

The input is now on a separate SMA-cable direct to the transformer. Repeat: no amplifier and no filter in this signal path. The sensitivity is 15 dB less than the original dongle without this modification. The S-meter reading minus 13.6 dB corresponds with the signal generator dBm value. The minimum discernible signal is now -90 dBm. So the amplification lacks about 15 to 20 dB; that equals the amplification of the removed amplifier.

There is, after this modification, no input bandwidth limitation. I added (externally) a 14 MHz low pass ( two inductors of 1 uH, input and output capacitance 150 or 180 pF and in the middle 300 or 330 pF to ground). And 12 pF across the first inductor and 33 pF across the second inductor. See attachement for diagram and simulator response.

The input impedance is NOT 50 ohm any more. This modified dongle is INTENDED to use with 50 ohm source. The measured sensitivity (over frequency) and frequency response are almost perfect when measured with a 50 ohm source impedance. The actual input impedance at the input of the transformer is higher, roughly 600 to 800 ohm real parallel 77 uH parallel 150 pF.
*** Added jan 31: real part is lower; closer to 400 Ohm) ***

When using this filter with the MODIFIED dongle (higher input impedance) then it is advised to match the output by insertion of a 50 ohm attenuator of a few dB, or just a 100 Ohm resistor across the output of the filter (is input of transformer). And then connect it to the dongle input: direct to the modified transformer.

Advise for followers of the modification: replace the BGA amplifier (NEC has identical amplifier, but an inductor must be added from output pin to plus: no latch-up) and replace the transformer. Be prepared for extreme miniature work: first check that you can produce such a transformer; otherwise buy it. Check if that transformer fits in the elegant housing.

The original V3 dongle design has IMHO no further shortcomings.

Dongle-LPF.jpg (54.84 KiB) Viewed 3292 times

Post Reply