I recently purchased a restock of CD4504 level converter chips for my AX80 testbed. Bascially these chips take in the 5V TTL signals and output 3.3v signals to the CMOS 74HC165 chips and ESP32 processors. Previous interations and I’ve built 3 work perfectly, but with this latest batch of 20 the outputs are all high regardless of the inputs switching. I also tested the chips in my cheap chip tester and they all pass the test for a CD4504. I tried the select lines high and low for compatability but still no joy. My only concern is the input voltage is 4V and dips perfectly to 0V, but why do my other chips all work and this batch do not. Good chip on the left, bad chip on the right, both TI badged.
If you haven’t already, I’d consider wiring it up on a bread board and seeing if you can make it work completely out of your circuit. Try with a bunch of values for the input and supply voltages (I see you mentioned 4V, though not sure if that’s the input voltage or which supply voltage you’re referring to) and see how things go. It’s possible that you’ve probably already done this, but I thought it best to mention the process in case anyone else stumbles across a similar issue.
One thing I will note is that the TI datasheet notes on page 3 of the PDF (https://www.ti.com/lit/ds/symlink/cd4504b.pdf) that Vdd (output supply) should be between 5V and 18V for reliable operation, and just under that all the examples in the following table (in the SHIFTING MODE column) all have both Vcc and Vdd being 5V or more. It’s possible that some of the newer CD4050B chips actually now handle lower Vdd voltages through changes to the underlying fab processes. If the chips on the left are new stock done on a newer fab process than the ones on the right (eg: right are NOS), then this could explain differences in operation.
If that’s the case, then this might mean you’re technically out of spec, and the other chips just “happen” to work in your use case. This is not great for repeatability and means you end up having to do your own binning of parts, ie: checking they work to your satisfaction and then getting rid of the ones that don’t meet your specs.
FWIW: For shifting down to 3.3V, IMO you probably want a 74HC4050, tho you lose the SELECT input (which determines if you’re doing TTL to CMOS or CMOS to CMOS anyway, which might not be necessary). They’re almost pin for pin compatible (re: inputs and outputs). The Vdd and SELECT pins are marked N.C, so you’d need to make sure they’re not connected (just in case they’re somehow actually linked within the plastic carrier to something). If they’re in sockets, you could just bend those pins up as a test to see if it’d work before committing to a design change/board mod.
Back to the CD4504B. Assuming that they still don’t work in any test circuit, the one pictured on the right might be fake. The fact that the code ABOVE the part number is different (2-part code with a 2 character suffix, vs string of numbers) is probably a clue. As said though, this might just be new stock with a longer code vs old stock with a shorter one.
I’d also look at the bottom of the chip. Many IC’s have plastic mould markings on the bottom that indicate the mould that was used to press the plastic casing. If the bottom of the chips that are not working look like they’ve been sanded or don’t match the working ones on the bottom, then you might have some fakes. Some links that might be useful reading for trying to determine if they are fakes:
https://www.aeri.com/counterfeit-electronic-component-detection/
https://www.allaboutcircuits.com/news/how-to-spot-counterfeit-electronic-components/
If you do find what could be a fake, and you bought these from a reliable/large distributor (eg: Digikey, E14, Mouser, etc), then I’d definitely be taking it up with them.
1 Like
I’m surprised by your answer, first the 4050 cannot level shift down, it’s quite capable of level shifting up as long as you pass the input thresholds for switching, I would never consider using a 4050 in any other way than as a buffer or maybe a 3.3v to 5v shifter. The 4504 however can shift up or down quite happily from 3.3 to 5 and vice versa if wired correctly. I have used 4504 chips in many projects over the last 4 years where it’s been required to provide 5v from a 3.3v microprocessor, the fact that I said I’ve built 3 of these circuits already and they all work with good chips should indicate that they will run at 3.3v. the datasheet says max -0.5 to 20v for VDD. 5, 10 and 15 are used as examples which is typical of data sheets.
I agree I should test them on a breadboard just to see if they actually do anything, which I will probably do later on. I could even test them in another circuit I have laying around where they are used for level shifting PWM signals to a sample and hold. This is a proven design and starts at 3.3v input and outputs 10v. I took 6 chips off this board to replace the non functioning ones in my latest build.
I just breadboarded them, 5V pin 1 and 3.3v pin 16. 0v on pin 3 (input) 3.3v on pin 2 output, 5V on pin 3 (input) and 3.3v on pin 2 (output). So no change. Test on pins 4 and 5, 0V on pin 4 regardless of pin 5 level.
Repeated the same test on a known good chip and everything works as expected.
Considering they only cost me 250 hrv for 20, so thats about 5 euros is it worth taking them back which would cost me more in fuel.
Note that I’m specifically talking about a 74HC4050, which is not the same thing as a CD4050.
Direct from an NXP datasheet for the 74HC4050 (right on page 1):
The 74HC4050 is a hex buffer with over-voltage tolerant inputs. Inputs are over-voltage
tolerant to 15 V which enables the device to be used in HIGH-to-LOW level shifting
applications.
And from the CD74HC4050 datasheet from TI (in the features):
- High-to-Low Voltage Level Converter for up to Vl = 16V
Note: HCT variants won’t do below 4.5V for the Vcc level.
So the 74HC4050 is specifically a clamping circuit to level shift down, the CD4504 will go up or down depending on VDD and VCC configurations, its a much more vesatile chip when it works.
1 Like
My local supplier was happy to take them back and refund me and inform his general manager that there is a problem with these chips.
2 Likes