Can't determine what I'm doing wrong in trying to scope serial com

[Nonyaz]

I bought a cheap WiFi chip that communicates via serial ESP-8266, and it wasn't initially responding to commands sent via the USB to serial device so I tried to hook up my scope to figure out what was or wasn't going on.  Tested the serial TX of the USB converter without connecting it to WiFi chip, looked fine.  Tested again with it connected to chip and it appears like its tied really hard to high, I adjusted my trigger and was able to see the data, but it wasn't reaching anywhere close to 0 like it was unconnected (see screenshot).

I eventually got the device to work without the help of the scope, and can now confirm that whenever I try to tap the TX line with the scope, the WiFi chip can't recognise the input, putting me into a quantum physics-esque situation where the act of monitoring it changes the outcome.

Undoubtedly I am doing something wrong to make this happen and so I will post some pictures of the setup and hopefully the problem will be obvious to someone.

[Gixy]

Recalibration of the scope? Same thing on another channel?

[Nonyaz]

Recalibration of the scope? Same thing on another channel?

Tried all channels.  Haven't calibrated the scope, I don't even think I have the equipment for that, but I don't know how that would account for the chip not being able to read the data anymore.

[bobcat]

Maybe it is a floating ground problem. Try using differential probing. (A-B)

[Gixy]

Calibration of the scope is an internal function, accessible under the "utility" menu. I agree that this should not affect the input signal. Bad reference should cause the issue, but your setup seems correct, except if the little breadboard or a cable has a defect.

[mij59]

Hi,

Don't it has anything to do with probing the circuit, looks to me like two outputs are connected together.
Maybe it's caused by poor power supply decoupling, put a low esr capacitor across the power supply pins of the ESP8266-05.

 

[vk6zgo]

Maybe it is a floating ground problem. Try using differential probing. (A-B)

I would think the same.

It may be that the Tx line in the first check was of a type which didn't mind being unbalanced when in the original condition,but when the WIFI chip was connected,this was no longer the case.
I've usually had to "fiddle about" to get a useable display when looking at serial signals,like RS-232 .etc but this may be different.

[AndyC_772]

Your scope isn't an a 50 Ohm input mode, is it?

What happens if, say, you keep the scope set up the same way, but probe across a AA battery? Does it read a sensible voltage then?

What if you then put a 10k resistor between the battery + terminal and the tip of the probe? Does it read the same, or something wildly different?

[jboard146]

Some things I noticed based on the pictures.

That's not the normal probes for the ds4k. Make sure you have 1x/10x configured. On those probes and in the scope.

Is the rs232 signal TTL 5v/3.3v or the "real rs232" ~15v

[Nonyaz]

Sorry for leaving you guys hanging for a while, I didn't get the notification that someone had responded.

Hi,

Don't it has anything to do with probing the circuit, looks to me like two outputs are connected together.
Maybe it's caused by poor power supply decoupling, put a low esr capacitor across the power supply pins of the ESP8266-05.

Tried the cap, same.

Your scope isn't an a 50 Ohm input mode, is it?

What happens if, say, you keep the scope set up the same way, but probe across a AA battery? Does it read a sensible voltage then?

What if you then put a 10k resistor between the battery + terminal and the tip of the probe? Does it read the same, or something wildly different?

I made the 50 Ohm mistake before on a different project (I really think they should hide that more, or have a confirmation box or something, its dangerous...) but no, that was one of the first things I checked.  Battery test works as expected.

Maybe it is a floating ground problem. Try using differential probing. (A-B)

I don't have a differential probe, unless you mean just use two and then use math, but even without connecting the ground it starts screwing up, I can't imagine how connecting another channel to another pin will reverse that, but if you insist I could try that.

Some things I noticed based on the pictures.

That's not the normal probes for the ds4k. Make sure you have 1x/10x configured. On those probes and in the scope.

Is the rs232 signal TTL 5v/3.3v or the "real rs232" ~15v

Good eye, I sold the probes on eBay and bought cheaper ones because I wanted 1x and didn't need high bandwidth capabilities (probably not my smartest move because I think I pretty much gave them away but thats another story).  The scope is set for 1x just like the probe, and 10x doesn't change anything.

The RS323 is 3.3V, I've never heard of 15V, but nothing like that is coming out of my USB port.


This house may or maynot have grounding issues (it's just some cheap student apartment) but I thought the ground line is independent from earth ground?

[Monkeh]

Do us a favour and link us to the exact serial and wifi boards you're using.

[Nonyaz]

http://www.ebay.com/itm/271722194511?_trksid=p2059210.m2749.l2649&ssPageName=STRK%3AMEBIDX%3AIT

Also this might be helpful: http://electro.olganet.com/wp-content/uploads/2014/11/esp8266-051.jpg

[Monkeh]

Disconnect the ESP8266 and see if TXD drives low properly when pulled up with, say, 2.2k.

[Electro Fan]

The RS323 is 3.3V, I've never heard of 15V, but nothing like that is coming out of my USB port.

Some differences between RS-232, TTL, and USB -

RS-232 vs. TTL Serial Communication:
https://www.sparkfun.com/tutorials/215

"Serial communication at a TTL level will always remain between the limits of 0V and Vcc, which is often 5V or 3.3V. A logic high ('1') is represented by Vcc, while a logic low ('0') is 0V.

By the RS-232 standard a logic high ('1') is represented by a negative voltage – anywhere from -3 to -25V – while a logic low ('0') transmits a positive voltage that can be anywhere from +3 to +25V. On most PCs these signals swing from -13 to +13V.

The more extreme voltages of an RS-232 signal help to make it less susceptible to noise, interference, and degradation. This means that an RS-232 signal can generally travel longer physical distances than their TTL counterparts, while still providing a reliable data transmission."

USB:
http://www.ti.com/lit/an/slyt118/slyt118.pdf
http://www.beyondlogic.org/usbnutshell/usb2.shtml

[Nonyaz]

Disconnect the ESP8266 and see if TXD drives low properly when pulled up with, say, 2.2k.

I'm not sure how to measure this properly while isolating my scope as a variable, but with a 2.7k, it pulls down 2.2V from peak aka down to 1.1V according to my scope.  Not sure what to conclude from this result..  Say I determined it wasn't pulling low hard enough, but why would adding a scope suddenly put it over the edge?

[Monkeh]

Disconnect the ESP8266 and see if TXD drives low properly when pulled up with, say, 2.2k.

I'm not sure how to measure this properly while isolating my scope as a variable, but with a 2.7k, it pulls down 2.2V from peak aka down to 1.1V according to my scope.  Not sure what to conclude from this result..  Say I determined it wasn't pulling low hard enough, but why would adding a scope suddenly put it over the edge?

Could be as simple as the extra capacitance pushing it over the edge.

With 2.7k pullup the UART should be able to pull it right the way down. It's meant to have 4mA sink capability. Either the datasheet lies (would not surprise me) or the chip is faulty (also would not surprise me).

Try another UART. Do the pullup test first, virtually anything should sink that.

[mij59]

Hi,

Try a pull up resistor on the RST pin of the ESP8266

[tautech]

Some things I noticed based on the pictures.

That's not the normal probes for the ds4k. Make sure you have 1x/10x configured. On those probes and in the scope.

Is the rs232 signal TTL 5v/3.3v or the "real rs232" ~15v

Good eye, I sold the probes on eBay and bought cheaper ones because I wanted 1x and didn't need high bandwidth capabilities (probably not my smartest move because I think I pretty much gave them away but thats another story).  The scope is set for 1x just like the probe, and 10x doesn't change anything.

YOU DID WHAT? 
Are you not aware of the many times additional capacitive loading you place on a DUT by using a 1x probe?

The scope is set for 1x just like the probe, and 10x doesn't change anything.

You can NO LONGER trust what you see on your scope over a few tens of KHz UNTIL you have 10x probes.
Some small homework for you: http://www.tek.com/document/primer/abcs-probes

[Nonyaz]

You can NO LONGER trust what you see on your scope over a few tens of KHz UNTIL you have 10x probes.
Some small homework for you: http://www.tek.com/document/primer/abcs-probes

The probes are 1x/10x so I do have 10x, I just meant to say that whether it was on 1x or 10x didn't change anything.  If you so wish, I would happily accept your donation of a better probe(s)

With 2.7k pullup the UART should be able to pull it right the way down. It's meant to have 4mA sink capability. Either the datasheet lies (would not surprise me) or the chip is faulty (also would not surprise me).

Try another UART. Do the pullup test first, virtually anything should sink that.

I'm starting to think this is the problem.  I swear I tested this before, but I put the scope on the TX pin of the ESP8266 and it is pulling down all the way to 0, previously I thought I had tested this and was unable to get a reply back so I would never see anything.  I do have another UART somewhere...  I'll try to find it and hope the drivers install automatically because I have no idea what brand it is anymore.