We just notcied a strange behavior of Acrobat/Adobe Reader (DC 2019.021.20056) and found the root cause:

There's a generated PDF that uses standard fonts which are referenced through the AcroForm default resources entry, too. Strangely the font has a Name entry in its dictionary:

3 0 obj
<</Type/Font/Subtype/Type1/BaseFont/Helvetica/Name/F1/Encoding/WinAnsiEncoding>>
endobj

We add 2 signature fields with our own tool and certify/sign one field. This results in a valid document, which you can download here (https://customers.setasign.com/adobe/document-with-name-in-font.pdf). Do not wonder, there's really only a single line on the page.

When you open this document in Adobe Acrobat or Reader and sign the left field the new signature is fine but the previous one is invalid because of changes which were not allowed. The resulting document is available here (https://customers.setasign.com/adobe/document-with-name-in-font-signed.pdf).

We tracked it down that Adobe Acrobat/Reader simply removes the Name entry from the font dictionary which invalidates the first signature:

3 0 obj
<</BaseFont/Helvetica/Encoding/WinAnsiEncoding/Subtype/Type1/Type/Font>>
endobj

If we remove the Name key before we add the signature fields and the first signature everything runs without any issue.

So from my point of view this seems to be a bug in Acrobat, or? Not to say that the generation library should omit the Name entry, too.