Drift on R820T

Status
Not open for further replies.

cellphone

Silent key.
Premium Subscriber
Joined
Dec 28, 2002
Messages
1,811
Location
Ahwatukee, AZ (Phoenix)
I recently purchased two RTL2832U/R820T SDRs from NooElec. These are fascinating devices that open up a whole new world for radio monitoring.



My problem is drift. I primarily use them to feed control channels into Unitrunker. Every few minutes, the signal will drop down to nothing, and Unitrunker will stop decoding. I adjust the correction by a couple ppm and it is back up to full decoding again. I tried using one to follow digital voice with DSD+, but it was frustrating because once I got good decode it would get bad because the receiver had drifted. I adjust the correction, and I’m back to good audio until it drifts again.



Things get better after they have been running for about 15 mins, but even then they will drift. Am I missing some configuration or setting that will help me with my drift issue?
 
Last edited:

BM82557

Member
Joined
Aug 28, 2006
Messages
4,972
Location
Berkeley Co WV
I had that problem also. I found that I was off 2 ppm on the correction. That along with turning on Autogain corrected it for me.
 

Token

Member
Joined
Jun 18, 2010
Messages
2,379
Location
Mojave Desert, California, USA
All of my RTL-SDRs drift, R820Ts and E4000s. I think it is just the nature of the beast.

Remember these were not mean to be SDRs used for relatively narrow banded radio signals as hobbyist are now using them. They were meant (designed) to work with digital TV signals, much more forgiving in frequency tolerance and including their own full time reference (pilot) tone transmission.

T!
 

cellphone

Silent key.
Premium Subscriber
Joined
Dec 28, 2002
Messages
1,811
Location
Ahwatukee, AZ (Phoenix)
That's what I was thinking considering it is a $20 device. I'm very interested in the HackRF One, but that is a lot pricier at $300+ each. Can anyone comment on on drift problems with these?
 

rbm

Member
Joined
Jan 25, 2005
Messages
1,395
Location
Upstate New York
Yes, they do drift quite a bit from a cold start.

However, after about 20-30 minutes or so, they settle down.

Once mine warmed up, I calibrated each (using separate SDR# folders).
(I have 10 dongles and they range from 17-89 ppm)
From that point on, they stay within around 1 ppm which I've found is sufficient for all digital modes I've received.
If that drift is too much for what you're doing, you can increase the bandwidth slightly.

Edit: This how I calibrate mine. http://forums.radioreference.com/software-defined-radio/294056-frequency-correction.html#post2213388

Rich

Here's what the cold start warm up looks like in SDRSharp
After another ten minutes or so, it's right on frequency and stays there. Mostly ;)

 
Last edited:

MileHighAko

Member
Joined
Aug 12, 2008
Messages
10
Location
Washington County, Utah
That's what I was thinking considering it is a $20 device. I'm very interested in the HackRF One, but that is a lot pricier at $300+ each. Can anyone comment on on drift problems with these?







HackRF also drifts, but doesn't seem as bad as the rtl-sdr, but still requires frequency correction. In both cases we're talking about inexpensive crystals.



I suspect the proper way to solve this problem is though smarter software that senses and autocorrects drifting problems.
 

joeuser

The Wretched
Joined
Feb 9, 2014
Messages
1,613
Location
North Central Kansas
I have 2 NooElec Dongles and they will Not stay on frequency.
They are supposed to be the best, but I disagree.
I do also, they work great. If they get cool, it takes a bit to get them warmed up. Once they are - they are good until the next power outage. If you are having issues I would look at the environment... Vents, windows, etc that can cause them to get hot or cold.
 

SCPD

QRT
Joined
Feb 24, 2001
Messages
0
Location
Virginia
Even after warm-up, the air around the RTL stick needs to be a consistent temperature. If you don't have A/C - or have a window A/C unit that alternates between blasting cold air and no air - that will cause drift.
 

mm

Member
Joined
Dec 19, 2002
Messages
659
Location
oregon
wow

I have a similar RTL but mine is a E4000 model. it seems that the no-elec models are junk.

I can monitor my Agilent RF generators output, set to -20dbm at 150MHz with the Agilent locked to a 10 MHz reference, and after ~4 minutes warm up time it never moves more than 1 ppm after several hours of use.

I was thinking of changing to the No-Elect 820 model that operates down to 25MHz but now I think I'll make a tcxo locked HF converter instead.
 

rbm

Member
Joined
Jan 25, 2005
Messages
1,395
Location
Upstate New York
The NooElec dongles and many of the Chinese dongles are identical.
There is nothing magical about them.

I've found the R820T tuners to be much better than the E4000 versions.
The only benefit with the E4000 tuners is they will tune a little higher in frequency.

Elonics E4000 52 - 2200 MHz with a gap from 1100 MHz to 1250 MHz (varies)
Rafael Micro R820T 24 - 1766 MHz

If you need to cover above 1766 MHz, go for the E4000.
Otherwise, the R820T is much better.

I've had around 20 dongles so far. I gave away half of them to friends.

I still have at least eight with the R820T tuners and 2 with the E4000 tuners.

I almost never bother using the E4000 tuners because the R820T is so much better.

Rich

Edit: By the way, I did buy 4 from NooElec so I know first hand how they compare.
Since they're in the US, you can get them much quicker. But other than that, I see no benefit.
 

cellphone

Silent key.
Premium Subscriber
Joined
Dec 28, 2002
Messages
1,811
Location
Ahwatukee, AZ (Phoenix)
Even after warm-up, the air around the RTL stick needs to be a consistent temperature. If you don't have A/C - or have a window A/C unit that alternates between blasting cold air and no air - that will cause drift.

At risk of going off topic, I just want to thank Unitrunker for the amazing software. I have learned so much about radio systems from this software. I used to live in an area covered by a Provoice system. If DSD would have been around then, in conjunction with Unitrunker, I would have been in hog heaven. I cannot get into software development, but I have great respect of those of you that can. With the software you create, you make it seem so easy! I know it is a lot of work. Adapting the software to support these SDRs is no exception. Keep up the great work for the scanner community!

Ok, back to drift discussion. :)
 

Token

Member
Joined
Jun 18, 2010
Messages
2,379
Location
Mojave Desert, California, USA
wow

I have a similar RTL but mine is a E4000 model. it seems that the no-elec models are junk.

I can monitor my Agilent RF generators output, set to -20dbm at 150MHz with the Agilent locked to a 10 MHz reference, and after ~4 minutes warm up time it never moves more than 1 ppm after several hours of use.

I was thinking of changing to the No-Elect 820 model that operates down to 25MHz but now I think I'll make a tcxo locked HF converter instead.

Nooelec did not have their R820's designed for them, they are the same as several other vendors just with custom silk screened cases with their name on the outside. They do not perform differently than other RTL SDRs.

I have several different RTL SDRs here, including both R820's and E4000's. I find they ALL drift with internal and external temperatures. If you keep them in a very stable (temperature) environment then after 15 or 20 minutes they drift very slowly. But drift they all do.

When you say "it never moves more than 1 ppm after several hours of use" I assume you mean you do not have to adjust the software calibration by more than 1 ppm. Because actually stability of 1 ppm at 150 MHz on the receiver would mean less than 150 Hz of total drift (less than +/- 75 Hz) in a couple of hours, and that exceeds the specification for the Icom R8500 at 150 MHz. Not one of my RTL SDRs come close to that kind of stability, regardless of make or model.

T!
 

moonbounce

Member
Joined
Nov 5, 2002
Messages
1,397
I just plugged my cold R820T in and ran it with SDR and DSD+. The r820t was cold to the touch, and when scanning, it wouldn't stop on anything. The only way to get it to stop on anything was to change the scan step to 500 Hz. About ten minutes later when the R820T had what I consider warmed up to operating temperature, it was hitting everything in SDR with a spike. I go through this procedure anytime I have not been using the dongle, even though it is always plugged in to the USB port.

Moonbounce
 

rbm

Member
Joined
Jan 25, 2005
Messages
1,395
Location
Upstate New York
Depending on the hardware settings, the USB port may be shut down during periods of inactivity.

Then, when you begin to use the port, the dongle will have to warm up all over again.

Rich

USB Power Management

 

BM82557

Member
Joined
Aug 28, 2006
Messages
4,972
Location
Berkeley Co WV
Depending on the hardware settings, the USB port may be shut down during periods of inactivity.

Then, when you begin to use the port, the dongle will have to warm up all over again.

Rich

USB Power Management



Thanks for that info, mine were set (Windows default??) to power off.
 

moonbounce

Member
Joined
Nov 5, 2002
Messages
1,397
Depending on the hardware settings, the USB port may be shut down during periods of inactivity.

Then, when you begin to use the port, the dongle will have to warm up all over again.

Rich

USB Power Management


Thanks Rich, my USB ports were also set to shut off when not active.

Moonbounce
 

Project25_MASTR

Millennial Graying OBT Guy
Joined
Jun 16, 2013
Messages
4,164
Location
Texas
One of my friends on Arfcom puts his RTL dongles in a small tin filled with mineral oil. He says that helps out a lot. I figure I may do that with one of mine in a pint sized paint can using some mineral oil and a bulkhead adapter.

I had a great idea of using the little dongles as repeater receivers…the drift issue stopped me before I really got the idea up on a tower.
 

slicerwizard

Member
Joined
Sep 19, 2002
Messages
7,643
Location
Toronto, Ontario
I suspect the proper way to solve this problem is though smarter software that senses and autocorrects drifting problems.
Bingo. TRUNK88 users never have this problem. It tracks drifts of 10, 20, whatever kHz. I used to be able to tell when the cat decided to take a nap on the dongles as the RX would crap out, but no more. Stupid cat...

Other software should be able to do the same; it's just a matter of implementing the logic.
 

moonbounce

Member
Joined
Nov 5, 2002
Messages
1,397
Thanks Rich, my USB ports were also set to shut off when not active.

Moonbounce

I changed the USB drives to be continually on, and I ran the R820T after being out of SDR for a couple of hours and no drift. It hit every frequency as if it was running for hours. Thanks again Rich.

Moonbounce
 
Status
Not open for further replies.
Top